Compare commits

...

13 Commits

Author SHA1 Message Date
jasonwitty f28635d6ad fix(deps): bump aws-lc-rs 1.15.0 -> 1.17.1 to patch aws-lc-sys advisories
CI / build (ubuntu-latest) (push) Has been cancelled
CI / build (windows-latest) (push) Has been cancelled
Resolves all five open Dependabot alerts (GHSA-9f94-5g5w-gf6r,
GHSA-394x-vwmw-crm3, GHSA-hfpc-8r3f-gw53, GHSA-65p9-r9h6-22vj,
GHSA-vw5v-4f2q-w9xf) by moving aws-lc-sys from 0.33.0 to 0.42.0.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-14 15:07:13 -07:00
jason 1c0c44ec3c fix(agent): gate now binding to logging feature
The Windows CI matrix surfaced an `unused_variables` warning at
metrics.rs:846 — `let now = std::time::Instant::now();` was bound
unconditionally but only consumed inside a `#[cfg(feature = "logging")]`
tracing::debug! call.

This block lives in the non-Linux `collect_processes_all`, so the Linux
CI never compiles it and never sees the warning. Same shape as the
`processes_ttl_ms` Windows fix from earlier: a binding whose only
consumer is cfg-gated needs to be cfg-gated too.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-06-03 11:15:45 -07:00
jason ee4468ca23 Add Debian packaging support for socktop-agent (#25)
* Add Debian packaging support with cargo-deb

- Add cargo-deb metadata to socktop and socktop_agent Cargo.toml
- Create systemd service file for socktop_agent
- Add postinst/postrm maintainer scripts for user/group management
- Create GitHub Actions workflow to build .deb packages for AMD64 and ARM64
- Add comprehensive documentation in docs/DEBIAN_PACKAGING.md
- Packages will be available as artifacts on every push
- Automatic GitHub releases for version tags

* Add summary documentation for debian packaging

* fix unit test, move to macro cargo_bin!

* hotfix for issue with socktop agent not creating ssl certificate on first launch after upgrade of axum server version.

* Add helpful post-install message to guide users on enabling socktop-agent service

* Fix CI build by installing libdrm development dependencies

* Fix package rename script - cargo-deb already includes architecture in filename

* Make GPU support optional to enable RISC-V builds without libdrm

- Add 'gpu' feature flag (enabled by default)
- Make gfxinfo dependency optional
- Provide no-op GPU metrics when gpu feature disabled
- Disable GPU support for RISC-V builds in CI (libdrm unavailable)
- All other architectures (amd64, arm64, armhf) still get GPU support

* feature gate GPU stats for arm v7

* specify correct package names.

* install aarch64-linux-gnu-gcc build dep

* specify correct package name

* add RISC-V GCC compiler

* add .cargo to gitignore to elimicate issue with riscv64-linux-gnu-gcc linker in config.toml

* add gcc-arm-linux-gnueabihf linker fore armv7

* set correct x-compile lib gcc-aarch64-linux-gnu for arm64 builds.

* add ports.ubuntu.com to sources

* Add ARM64 as a foreign architecture

* fixe for ARM64 build.

* security.ubuntu.com` aNNOYING

* apt repo github page

* copy output to apt repo

* fix secrets path

* fix secrets path

* change build dep

* Fix postinst message box alignment

* ci(deb): restrict APT publish to v* release tags

Previously the workflow built and published on every push to master
and feature/debian-packaging in addition to v* tags. That meant the
gh-pages APT repo got overwritten on every commit with same-version
.debs, causing apt clients to see a phantom "update available" each
time and burning ~5-10 min of cross-compile CI per merge.

After this change:
  - PRs into master still cross-build .debs as a sanity check.
  - v* tags build, publish to gh-pages, and create a GitHub release.
  - workflow_dispatch remains as the manual escape hatch.
  - master pushes no longer trigger this workflow (ci.yml still runs).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-06-02 13:12:51 -07:00
jason 697a77bdab chore: update ratatui from 0.28 to 0.30 (#33)
* chore: update ratatui from 0.28 to 0.30

* style: cargo fmt

* fix: replace manual zero-guarded divisions with checked_div

* fix: collapse nested if into match guard

* style: cargo fmt

* bump crossterm and optimize various types, remove stale code.

* fix windows build

* only show parent level processes on main tui
2026-06-02 13:02:25 -07:00
dependabot[bot] 8f452a35e6 Bump rand from 0.8.5 to 0.8.6 (#31)
Bumps [rand](https://github.com/rust-random/rand) from 0.8.5 to 0.8.6.
- [Release notes](https://github.com/rust-random/rand/releases)
- [Changelog](https://github.com/rust-random/rand/blob/0.8.6/CHANGELOG.md)
- [Commits](https://github.com/rust-random/rand/compare/0.8.5...0.8.6)

---
updated-dependencies:
- dependency-name: rand
  dependency-version: 0.8.6
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-06-02 13:01:57 -07:00
dependabot[bot] 1d285c3c4e Bump rustls-webpki from 0.103.10 to 0.103.13 (#32)
Bumps [rustls-webpki](https://github.com/rustls/webpki) from 0.103.10 to 0.103.13.
- [Release notes](https://github.com/rustls/webpki/releases)
- [Commits](https://github.com/rustls/webpki/compare/v/0.103.10...v/0.103.13)

---
updated-dependencies:
- dependency-name: rustls-webpki
  dependency-version: 0.103.13
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-06-02 13:01:30 -07:00
dependabot[bot] 0fb45f6c50 Bump rustls-webpki from 0.103.8 to 0.103.10 (#30)
Bumps [rustls-webpki](https://github.com/rustls/webpki) from 0.103.8 to 0.103.10.
- [Release notes](https://github.com/rustls/webpki/releases)
- [Commits](https://github.com/rustls/webpki/compare/v/0.103.8...v/0.103.10)

---
updated-dependencies:
- dependency-name: rustls-webpki
  dependency-version: 0.103.10
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-04-13 09:43:38 -07:00
dependabot[bot] be24fa3859 Bump bytes from 1.10.1 to 1.11.1 in /socktop_wasm_test (#27)
Bumps [bytes](https://github.com/tokio-rs/bytes) from 1.10.1 to 1.11.1.
- [Release notes](https://github.com/tokio-rs/bytes/releases)
- [Changelog](https://github.com/tokio-rs/bytes/blob/master/CHANGELOG.md)
- [Commits](https://github.com/tokio-rs/bytes/compare/v1.10.1...v1.11.1)

---
updated-dependencies:
- dependency-name: bytes
  dependency-version: 1.11.1
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-02-25 16:14:57 -08:00
dependabot[bot] 39619b2845 Bump bytes from 1.11.0 to 1.11.1 (#28)
Bumps [bytes](https://github.com/tokio-rs/bytes) from 1.11.0 to 1.11.1.
- [Release notes](https://github.com/tokio-rs/bytes/releases)
- [Changelog](https://github.com/tokio-rs/bytes/blob/master/CHANGELOG.md)
- [Commits](https://github.com/tokio-rs/bytes/compare/v1.11.0...v1.11.1)

---
updated-dependencies:
- dependency-name: bytes
  dependency-version: 1.11.1
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-02-25 16:14:46 -08:00
dependabot[bot] 40c925e7f9 Bump time from 0.3.44 to 0.3.47 (#29)
Bumps [time](https://github.com/time-rs/time) from 0.3.44 to 0.3.47.
- [Release notes](https://github.com/time-rs/time/releases)
- [Changelog](https://github.com/time-rs/time/blob/main/CHANGELOG.md)
- [Commits](https://github.com/time-rs/time/compare/v0.3.44...v0.3.47)

---
updated-dependencies:
- dependency-name: time
  dependency-version: 0.3.47
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-02-25 16:14:29 -08:00
jason 8f69e469e6 Update README with website link 2025-11-30 10:21:26 -08:00
jason 3024816525 hotfix for issue with socktop agent not creating ssl certificate on first launch after upgrade of axum server version.
CI / build (ubuntu-latest) (push) Has been cancelled
CI / build (windows-latest) (push) Has been cancelled
2025-11-21 00:21:05 -08:00
jason 1d7bc42d59 fix unit test, move to macro cargo_bin! 2025-11-21 00:07:44 -08:00
43 changed files with 3271 additions and 1083 deletions
+422
View File
@@ -0,0 +1,422 @@
name: Build Debian Packages
on:
# APT publishing is release-driven: we build + publish only on `v*` tag
# pushes. PRs into master still build the .debs as a sanity check (no
# publish). Manual dispatch is kept as an escape hatch.
push:
tags:
- "v*"
pull_request:
branches:
- master
workflow_dispatch:
env:
CARGO_TERM_COLOR: always
jobs:
build-deb:
name: Build .deb for ${{ matrix.target }}
runs-on: ubuntu-latest
strategy:
matrix:
include:
- target: x86_64-unknown-linux-gnu
arch: amd64
- target: aarch64-unknown-linux-gnu
arch: arm64
- target: armv7-unknown-linux-gnueabihf
arch: armhf
- target: riscv64gc-unknown-linux-gnu
arch: riscv64
steps:
- name: Checkout code
uses: actions/checkout@v4
- name: Install Rust toolchain
uses: dtolnay/rust-toolchain@stable
with:
targets: ${{ matrix.target }}
- name: Install cargo-deb
run: cargo install cargo-deb
- name: Install build dependencies
run: |
sudo apt-get update
sudo apt-get install -y dpkg-dev
- name: Install cross-compilation tools (ARM64)
if: matrix.target == 'aarch64-unknown-linux-gnu'
run: |
sudo dpkg --add-architecture arm64
# Disable all existing sources and create new ones with proper arch specifications
sudo mv /etc/apt/sources.list /etc/apt/sources.list.backup
sudo mv /etc/apt/sources.list.d /etc/apt/sources.list.d.backup || true
sudo mkdir -p /etc/apt/sources.list.d
# Clear APT cache and lists
sudo rm -rf /var/lib/apt/lists/*
sudo mkdir -p /var/lib/apt/lists/partial
# Create new sources.list with both amd64 and arm64
cat << EOF | sudo tee /etc/apt/sources.list
deb [arch=amd64] http://archive.ubuntu.com/ubuntu $(lsb_release -sc) main universe restricted multiverse
deb [arch=amd64] http://archive.ubuntu.com/ubuntu $(lsb_release -sc)-updates main universe restricted multiverse
deb [arch=amd64] http://archive.ubuntu.com/ubuntu $(lsb_release -sc)-backports main universe restricted multiverse
deb [arch=amd64] http://security.ubuntu.com/ubuntu $(lsb_release -sc)-security main universe restricted multiverse
deb [arch=arm64] http://ports.ubuntu.com/ubuntu-ports $(lsb_release -sc) main universe restricted multiverse
deb [arch=arm64] http://ports.ubuntu.com/ubuntu-ports $(lsb_release -sc)-updates main universe restricted multiverse
deb [arch=arm64] http://ports.ubuntu.com/ubuntu-ports $(lsb_release -sc)-backports main universe restricted multiverse
deb [arch=arm64] http://ports.ubuntu.com/ubuntu-ports $(lsb_release -sc)-security main universe restricted multiverse
EOF
echo "=== Contents of /etc/apt/sources.list ==="
cat /etc/apt/sources.list
echo "=== Contents of /etc/apt/sources.list.d/ ==="
ls -la /etc/apt/sources.list.d/ || true
sudo apt-get update
sudo apt-get install -y gcc-aarch64-linux-gnu libdrm-dev:arm64 libdrm-amdgpu1:arm64
- name: Install cross-compilation tools (ARMhf)
if: matrix.target == 'armv7-unknown-linux-gnueabihf'
run: |
sudo apt-get update
sudo apt-get install -y gcc-arm-linux-gnueabihf
- name: Install cross-compilation tools (RISC-V)
if: matrix.target == 'riscv64gc-unknown-linux-gnu'
run: |
sudo apt-get update
sudo apt-get install -y gcc-riscv64-linux-gnu
- name: Install GPU libraries (x86_64)
if: matrix.target == 'x86_64-unknown-linux-gnu'
run: |
sudo apt-get update
sudo apt-get install -y libdrm-dev libdrm-amdgpu1
- name: Configure cross-compilation (ARM64)
if: matrix.target == 'aarch64-unknown-linux-gnu'
run: |
mkdir -p .cargo
cat >> .cargo/config.toml << EOF
[target.aarch64-unknown-linux-gnu]
linker = "aarch64-linux-gnu-gcc"
EOF
- name: Configure cross-compilation (ARMhf)
if: matrix.target == 'armv7-unknown-linux-gnueabihf'
run: |
mkdir -p .cargo
cat >> .cargo/config.toml << EOF
[target.armv7-unknown-linux-gnueabihf]
linker = "arm-linux-gnueabihf-gcc"
EOF
- name: Configure cross-compilation (RISC-V)
if: matrix.target == 'riscv64gc-unknown-linux-gnu'
run: |
mkdir -p .cargo
cat >> .cargo/config.toml << EOF
[target.riscv64gc-unknown-linux-gnu]
linker = "riscv64-linux-gnu-gcc"
EOF
- name: Cache cargo registry
uses: actions/cache@v4
with:
path: ~/.cargo/registry
key: ${{ runner.os }}-cargo-registry-${{ hashFiles('**/Cargo.lock') }}
- name: Cache cargo index
uses: actions/cache@v4
with:
path: ~/.cargo/git
key: ${{ runner.os }}-cargo-index-${{ hashFiles('**/Cargo.lock') }}
- name: Cache target directory
uses: actions/cache@v4
with:
path: target
key: ${{ runner.os }}-target-${{ matrix.target }}-${{ hashFiles('**/Cargo.lock') }}
- name: Build socktop .deb package
run: |
cargo deb --package socktop --target ${{ matrix.target }} --no-strip
- name: Build socktop_agent .deb package (with GPU support)
if: matrix.target == 'x86_64-unknown-linux-gnu' || matrix.target == 'aarch64-unknown-linux-gnu'
run: |
cargo deb --package socktop_agent --target ${{ matrix.target }} --no-strip
- name: Build socktop_agent .deb package (without GPU support)
if: matrix.target == 'armv7-unknown-linux-gnueabihf' || matrix.target == 'riscv64gc-unknown-linux-gnu'
run: |
cargo deb --package socktop_agent --target ${{ matrix.target }} --no-strip --no-default-features
- name: Copy packages to debs directory
run: |
mkdir -p debs
cp target/${{ matrix.target }}/debian/*.deb debs/
- name: List generated packages
run: ls -lh debs/
- name: Upload .deb packages as artifacts
uses: actions/upload-artifact@v4
with:
name: debian-packages-${{ matrix.arch }}
path: debs/*.deb
if-no-files-found: error
retention-days: 90
# Combine all artifacts into a single downloadable archive
combine-artifacts:
name: Combine all .deb packages
needs: build-deb
runs-on: ubuntu-latest
steps:
- name: Download AMD64 packages
uses: actions/download-artifact@v4
with:
name: debian-packages-amd64
path: all-debs
- name: Download ARM64 packages
uses: actions/download-artifact@v4
with:
name: debian-packages-arm64
path: all-debs
- name: Download ARMhf packages
uses: actions/download-artifact@v4
with:
name: debian-packages-armhf
path: all-debs
- name: Download RISC-V packages
uses: actions/download-artifact@v4
with:
name: debian-packages-riscv64
path: all-debs
- name: List all packages
run: |
echo "All generated .deb packages:"
ls -lh all-debs/
- name: Upload combined artifacts
uses: actions/upload-artifact@v4
with:
name: all-debian-packages
path: all-debs/*.deb
if-no-files-found: error
retention-days: 90
- name: Generate checksums
run: |
cd all-debs
sha256sum *.deb > SHA256SUMS
cat SHA256SUMS
- name: Upload checksums
uses: actions/upload-artifact@v4
with:
name: checksums
path: all-debs/SHA256SUMS
retention-days: 90
# Publish packages to gh-pages APT repository
publish-apt-repo:
name: Publish to APT Repository
needs: combine-artifacts
runs-on: ubuntu-latest
# Publish only on `v*` release tags — keep gh-pages stable between
# releases instead of overwriting same-version .debs on every commit.
if: startsWith(github.ref, 'refs/tags/v')
permissions:
contents: write
steps:
- name: Checkout code
uses: actions/checkout@v4
with:
fetch-depth: 0
- name: Download all packages
uses: actions/download-artifact@v4
with:
name: all-debian-packages
path: debs
- name: Install dependencies
run: |
sudo apt-get update
sudo apt-get install -y dpkg-dev gpg
- name: Checkout gh-pages branch
run: |
git fetch origin gh-pages:gh-pages || echo "gh-pages branch doesn't exist yet"
if git show-ref --verify --quiet refs/heads/gh-pages; then
git checkout gh-pages
else
git checkout --orphan gh-pages
git rm -rf . 2>/dev/null || true
# Create basic structure
mkdir -p dists/stable/main/{binary-amd64,binary-arm64,binary-armhf,binary-riscv64}
mkdir -p pool/main
fi
- name: Copy packages to pool
run: |
mkdir -p pool/main
cp debs/*.deb pool/main/
ls -lh pool/main/
- name: Generate Packages files
run: |
for arch in amd64 arm64 armhf riscv64; do
mkdir -p dists/stable/main/binary-$arch
dpkg-scanpackages --arch $arch pool/main /dev/null > dists/stable/main/binary-$arch/Packages 2>/dev/null || true
if [ -s dists/stable/main/binary-$arch/Packages ]; then
gzip -9 -k -f dists/stable/main/binary-$arch/Packages
echo "Generated Packages file for $arch"
fi
done
- name: Generate Release file
run: |
cat > dists/stable/Release << EOF
Origin: socktop
Label: socktop
Suite: stable
Codename: stable
Architectures: amd64 arm64 armhf riscv64
Components: main
Description: socktop APT repository
Date: $(date -Ru)
EOF
# Add MD5Sum
echo "MD5Sum:" >> dists/stable/Release
for arch in amd64 arm64 armhf riscv64; do
for file in dists/stable/main/binary-$arch/Packages*; do
if [ -f "$file" ]; then
md5sum "$file" | awk '{print " " $1, "'$(stat -c%s "$file" 2>/dev/null || stat -f%z "$file" 2>/dev/null)'", "'"${file#dists/stable/}"'"}' >> dists/stable/Release
fi
done
done
# Add SHA256
echo "SHA256:" >> dists/stable/Release
for arch in amd64 arm64 armhf riscv64; do
for file in dists/stable/main/binary-$arch/Packages*; do
if [ -f "$file" ]; then
sha256sum "$file" | awk '{print " " $1, "'$(stat -c%s "$file" 2>/dev/null || stat -f%z "$file" 2>/dev/null)'", "'"${file#dists/stable/}"'"}' >> dists/stable/Release
fi
done
done
- name: Set GPG available flag
id: check_gpg
env:
GPG_PRIVATE_KEY: ${{ secrets.GPG_PRIVATE_KEY }}
run: |
if [ -n "$GPG_PRIVATE_KEY" ]; then
echo "available=true" >> $GITHUB_OUTPUT
else
echo "available=false" >> $GITHUB_OUTPUT
fi
- name: Import GPG key
if: steps.check_gpg.outputs.available == 'true'
env:
GPG_PRIVATE_KEY: ${{ secrets.GPG_PRIVATE_KEY }}
run: |
echo "$GPG_PRIVATE_KEY" | gpg --batch --import
gpg --list-secret-keys
- name: Sign repository
if: steps.check_gpg.outputs.available == 'true'
env:
GPG_KEY_ID: ${{ secrets.GPG_KEY_ID }}
GPG_PASSPHRASE: ${{ secrets.GPG_PASSPHRASE }}
run: |
if [ -n "$GPG_PASSPHRASE" ]; then
echo "$GPG_PASSPHRASE" | gpg --batch --yes --no-tty --pinentry-mode loopback --passphrase-fd 0 \
--default-key "$GPG_KEY_ID" \
-abs -o dists/stable/Release.gpg dists/stable/Release
echo "$GPG_PASSPHRASE" | gpg --batch --yes --no-tty --pinentry-mode loopback --passphrase-fd 0 \
--default-key "$GPG_KEY_ID" \
--clearsign -o dists/stable/InRelease dists/stable/Release
else
gpg --batch --yes --no-tty --pinentry-mode loopback \
--default-key "$GPG_KEY_ID" \
-abs -o dists/stable/Release.gpg dists/stable/Release
gpg --batch --yes --no-tty --pinentry-mode loopback \
--default-key "$GPG_KEY_ID" \
--clearsign -o dists/stable/InRelease dists/stable/Release
fi
gpg --armor --export "$GPG_KEY_ID" > KEY.gpg
echo "✓ Repository signed"
- name: Create unsigned repository notice
if: steps.check_gpg.outputs.available == 'false'
run: |
echo "⚠️ Warning: GPG_PRIVATE_KEY not set. Repository will be UNSIGNED."
echo "⚠️ Add GPG secrets to sign the repository automatically."
echo "To add secrets: Settings → Secrets and variables → Actions → Repository secrets"
- name: Copy index.html if exists
run: |
git checkout ${{ github.ref_name }} -- index.html 2>/dev/null || echo "No index.html in source branch"
- name: Commit and push to gh-pages
run: |
git config user.name "GitHub Actions"
git config user.email "actions@github.com"
git add .
if git diff --staged --quiet; then
echo "No changes to commit"
else
COMMIT_MSG="Update APT repository"
if [[ "${{ github.ref }}" == refs/tags/* ]]; then
COMMIT_MSG="$COMMIT_MSG - Release ${{ github.ref_name }}"
else
COMMIT_MSG="$COMMIT_MSG - $(date -u +'%Y-%m-%d %H:%M:%S UTC')"
fi
git commit -m "$COMMIT_MSG"
git push origin gh-pages
echo "✓ Published to gh-pages"
fi
# Optional: Create a release with the .deb files if this is a tag
create-release:
name: Create GitHub Release
needs: combine-artifacts
runs-on: ubuntu-latest
if: startsWith(github.ref, 'refs/tags/v')
permissions:
contents: write
steps:
- name: Download all packages
uses: actions/download-artifact@v4
with:
name: all-debian-packages
path: release-debs
- name: Download checksums
uses: actions/download-artifact@v4
with:
name: checksums
path: release-debs
- name: Create Release
uses: softprops/action-gh-release@v1
with:
files: release-debs/*
draft: false
prerelease: false
generate_release_notes: true
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+9
View File
@@ -1,7 +1,16 @@
/target
.vscode/
/socktop-wasm-test/target
/.cargo/
# Documentation files from development sessions (context-specific, not for public repo)
/OPTIMIZATION_PROCESS_DETAILS.md
/THREAD_SUPPORT.md
# APT Repository - Safety: Never commit private keys!
*.asc
*-private.key
*-secret.key
gpg-private-backup.key
secring.gpg
# Note: Release.gpg, InRelease, and KEY.gpg (public) ARE safe to commit
Generated
+892 -279
View File
File diff suppressed because it is too large Load Diff
+2 -2
View File
@@ -24,8 +24,8 @@ serde_json = "1.0"
sysinfo = "0.37"
# CLI UI
ratatui = "0.28"
crossterm = "0.27"
ratatui = "0.30"
crossterm = "0.29"
# web server (remote-agent)
axum = { version = "0.7", features = ["ws"] }
+156
View File
@@ -0,0 +1,156 @@
# Debian Packaging Implementation Summary
## Overview
Successfully implemented Debian packaging for socktop using `cargo-deb`, with GitHub Actions automation for building packages for both AMD64 and ARM64 architectures.
## Branches Created
1. **`feature/debian-packaging`** - Main branch with debian packaging implementation
2. **`feature/man-pages`** - Separate branch for man pages work (to be researched further)
## What Was Added
### 1. Cargo.toml Updates
Both `socktop/Cargo.toml` and `socktop_agent/Cargo.toml` were updated with:
- `[package.metadata.deb]` sections
- Package metadata (maintainer, description, dependencies)
- Asset definitions (binaries, documentation)
- Systemd service configuration (agent only)
### 2. Systemd Service
**File**: `socktop_agent/socktop-agent.service`
- Runs as `socktop` user/group
- Listens on port 3000 by default
- Security hardening enabled
- Disabled by default (user must explicitly enable)
### 3. Maintainer Scripts
**Directory**: `socktop_agent/debian/`
- **`postinst`**: Creates `socktop` user/group, sets up `/var/lib/socktop` directory
- **`postrm`**: Cleanup on package removal/purge
### 4. GitHub Actions Workflow
**File**: `.github/workflows/build-deb.yml`
Features:
- Builds for both x86_64 and ARM64
- Triggered on:
- Push to `master` or `feature/debian-packaging`
- Pull requests to `master`
- Version tags (v*)
- Manual workflow dispatch
- Creates artifacts:
- `debian-packages-amd64`
- `debian-packages-arm64`
- `all-debian-packages` (combined)
- `checksums` (SHA256SUMS)
- Automatic GitHub releases for version tags
### 5. Documentation
**File**: `docs/DEBIAN_PACKAGING.md`
Comprehensive guide covering:
- Building packages locally
- Cross-compilation for ARM64
- Installation and configuration
- Using GitHub Actions artifacts
- Creating local APT repositories
- Troubleshooting
## Package Details
### socktop (TUI Client)
- **Binary**: `/usr/bin/socktop`
- **Size**: ~3.5 MB (x86_64)
- **Dependencies**: Auto-detected
### socktop_agent (Daemon)
- **Binary**: `/usr/bin/socktop_agent`
- **Service**: `socktop-agent.service`
- **User/Group**: `socktop` (created automatically)
- **State directory**: `/var/lib/socktop`
- **Size**: ~6.7 MB (x86_64)
- **Dependencies**: Auto-detected
## Testing
Both packages successfully built locally:
```
✓ socktop_1.50.0-1_amd64.deb
✓ socktop-agent_1.50.1-1_amd64.deb
```
Verified:
- Package contents (dpkg -c)
- Package metadata (dpkg -I)
- Systemd service file inclusion
- Maintainer scripts inclusion
- Documentation inclusion
## Usage
### For Users
Download pre-built packages from GitHub Actions artifacts:
1. Go to Actions tab
2. Select latest "Build Debian Packages" run
3. Download architecture-specific artifact
4. Install: `sudo dpkg -i socktop*.deb`
### For Developers
Build locally:
```bash
cargo install cargo-deb
cargo deb --package socktop
cargo deb --package socktop_agent
```
Cross-compile for ARM64:
```bash
rustup target add aarch64-unknown-linux-gnu
sudo apt install gcc-aarch64-linux-gnu libc6-dev-arm64-cross
cargo deb --package socktop --target aarch64-unknown-linux-gnu
```
## Next Steps
To get packages in official APT repositories:
1. **Short term**: Host packages on GitHub Releases (automated)
2. **Medium term**: Create PPA for Ubuntu users
3. **Long term**: Submit to Debian/Ubuntu official repositories
## Files Modified/Created
```
Modified:
socktop/Cargo.toml
socktop_agent/Cargo.toml
Created:
.github/workflows/build-deb.yml
docs/DEBIAN_PACKAGING.md
socktop_agent/socktop-agent.service
socktop_agent/debian/postinst
socktop_agent/debian/postrm
```
## Commit
```
532ed16 Add Debian packaging support with cargo-deb
```
## Resources
- [cargo-deb documentation](https://github.com/kornelski/cargo-deb)
- [Debian Policy Manual](https://www.debian.org/doc/debian-policy/)
- Full documentation in `docs/DEBIAN_PACKAGING.md`
+12 -2
View File
@@ -5,6 +5,8 @@ socktop is a remote system monitor with a rich TUI, inspired by top/btop, talkin
- Linux agent: near-zero CPU when idle (request-driven, no always-on sampler)
- TUI: smooth graphs, sortable process table, scrollbars, readable colors
[socktop.io](https://www.socktop.io)
<img src="./docs/socktop_demo.apng" width="100%">
---
@@ -51,15 +53,23 @@ exec bash # or: exec zsh / exec fish
Windows (for the brave): install from https://rustup.rs with the MSVC toolchain. Yes, youll need Visual Studio Build Tools. You chose Windows — enjoy the ride.
### Raspberry Pi / Ubuntu / PopOS (required)
### Raspberry Pi / Ubuntu / PopOS (required for GPU support)
Install GPU support with apt command below
**Note:** GPU monitoring is only supported on x86_64 and aarch64 (64-bit ARM) platforms. ARMv7 (32-bit) and RISC-V builds do not include GPU support.
For 64-bit systems with GPU support:
```bash
sudo apt-get update
sudo apt-get install libdrm-dev libdrm-amdgpu1
```
For ARMv7 (32-bit Raspberry Pi), build with `--no-default-features` to disable GPU support:
```bash
cargo build --release -p socktop_agent --no-default-features
```
_Additional note for Raspberry Pi users. Please update your system to use the newest kernel available through app, kernel version 6.6+ will use considerably less overall CPU to run the agent. For example on a rpi4 the kernel < 6.6 the agent will consume .8 cpu but on the same hardware on > 6.6 the agent will consume only .2 cpu. (these numbers indicate continuous polling at web socket endpoints, when not in use the usage is 0)_
---
+42
View File
@@ -0,0 +1,42 @@
-----BEGIN PGP PUBLIC KEY BLOCK-----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=AaPQ
-----END PGP PUBLIC KEY BLOCK-----
+38
View File
@@ -0,0 +1,38 @@
# socktop APT Repository
This repository contains Debian packages for socktop and socktop-agent.
## Adding this repository
Add the repository to your system:
```bash
# Add the GPG key
curl -fsSL https://jasonwitty.github.io/socktop/KEY.gpg | sudo gpg --dearmor -o /usr/share/keyrings/socktop-archive-keyring.gpg
# Add the repository
echo "deb [signed-by=/usr/share/keyrings/socktop-archive-keyring.gpg] https://jasonwitty.github.io/socktop stable main" | sudo tee /etc/apt/sources.list.d/socktop.list
# Update and install
sudo apt update
sudo apt install socktop socktop-agent
```
## Manual Installation
You can also download and install packages manually from the `pool/main/` directory.
```bash
wget https://jasonwitty.github.io/socktop/pool/main/socktop_VERSION_ARCH.deb
sudo dpkg -i socktop_VERSION_ARCH.deb
```
## Supported Architectures
- amd64 (x86_64)
- arm64 (aarch64)
- armhf (32-bit ARM)
## Building from Source
See the main repository at https://github.com/jasonwitty/socktop
+32
View File
@@ -0,0 +1,32 @@
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
Origin: socktop
Label: socktop
Suite: stable
Codename: stable
Architectures: amd64 arm64 armhf
Components: main
Description: socktop APT repository
Date: Sun, 23 Nov 2025 04:05:21 +0000
MD5Sum:
0bddefb2f13cb7c86cd05fe1ce20310f 1549 main/binary-amd64/Packages
674f0e552cbb7dc65380651a2a8d279e 799 main/binary-amd64/Packages.gz
SHA256:
babfbb4839e7fdfbc83742c16996791b0402a1315889b530330b338380398263 1549 main/binary-amd64/Packages
f8c48d0f7bf53eb02c6dbf5f1cdd046fe71b87273cf763c5bb2e95d9757a7a82 799 main/binary-amd64/Packages.gz
-----BEGIN PGP SIGNATURE-----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=IeBY
-----END PGP SIGNATURE-----
+14
View File
@@ -0,0 +1,14 @@
Origin: socktop
Label: socktop
Suite: stable
Codename: stable
Architectures: amd64 arm64 armhf
Components: main
Description: socktop APT repository
Date: Sun, 23 Nov 2025 04:05:21 +0000
MD5Sum:
0bddefb2f13cb7c86cd05fe1ce20310f 1549 main/binary-amd64/Packages
674f0e552cbb7dc65380651a2a8d279e 799 main/binary-amd64/Packages.gz
SHA256:
babfbb4839e7fdfbc83742c16996791b0402a1315889b530330b338380398263 1549 main/binary-amd64/Packages
f8c48d0f7bf53eb02c6dbf5f1cdd046fe71b87273cf763c5bb2e95d9757a7a82 799 main/binary-amd64/Packages.gz
+14
View File
@@ -0,0 +1,14 @@
-----BEGIN PGP SIGNATURE-----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=QoLf
-----END PGP SIGNATURE-----
@@ -0,0 +1,38 @@
Package: socktop
Version: 1.50.0-1
Architecture: amd64
Maintainer: Jason Witty <jasonpwitty+socktop@proton.me>
Installed-Size: 3459
Filename: pool/main/socktop_1.50.0-1_amd64.deb
Size: 1278940
MD5sum: 0215e178e306d9379669065e8c78582b
SHA1: 04e0416389f5cecd584fd1f6b3568711f2645eee
SHA256: 69eb04b1de48541c95950a97b16357fcd9c51ffaceb143f63de4a9d758fad297
Section: admin
Priority: optional
Homepage: https://github.com/jasonwitty/socktop
Description: Remote system monitor over WebSocket, TUI like top
socktop is a remote system monitor with a rich terminal user interface (TUI)
that connects to remote hosts running the socktop_agent over WebSocket. It
provides real-time monitoring of CPU, memory, processes, and more with an
interface similar to the traditional 'top' command.
Package: socktop-agent
Version: 1.50.2-1
Architecture: amd64
Maintainer: Jason Witty <jasonpwitty+socktop@proton.me>
Installed-Size: 6793
Filename: pool/main/socktop-agent_1.50.2-1_amd64.deb
Size: 1896272
MD5sum: 22e78d03e83dcf84d6ec4a009b285902
SHA1: 26a9f4fedfdba06a047044027223f2944cf72ba6
SHA256: 11922af475146f60347a9c52cff4bbce1ce524bdb4293b2c436f3c71876e17d5
Section: admin
Priority: optional
Homepage: https://github.com/jasonwitty/socktop
Description: Socktop agent daemon. Serves host metrics over WebSocket.
socktop_agent is the daemon component that runs on remote hosts to collect and
serve system metrics over WebSocket. It gathers CPU, memory, disk, network,
GPU, and process information that can be monitored remotely by the socktop TUI
client.
Binary file not shown.
@@ -0,0 +1,5 @@
Archive: stable
Component: main
Origin: socktop
Label: socktop
Architecture: amd64
+58
View File
@@ -0,0 +1,58 @@
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>socktop APT Repository</title>
<style>
body {
font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif;
max-width: 800px;
margin: 50px auto;
padding: 20px;
line-height: 1.6;
}
code {
background: #f4f4f4;
padding: 2px 6px;
border-radius: 3px;
}
pre {
background: #f4f4f4;
padding: 15px;
border-radius: 5px;
overflow-x: auto;
}
h1 { color: #333; }
h2 { color: #555; margin-top: 30px; }
</style>
</head>
<body>
<h1>socktop APT Repository</h1>
<p>System monitor with remote agent support for Linux systems.</p>
<h2>Adding this repository</h2>
<pre><code># Add the GPG key
curl -fsSL https://jasonwitty.github.io/socktop/KEY.gpg | sudo gpg --dearmor -o /usr/share/keyrings/socktop-archive-keyring.gpg
# Add the repository
echo "deb [signed-by=/usr/share/keyrings/socktop-archive-keyring.gpg] https://jasonwitty.github.io/socktop stable main" | sudo tee /etc/apt/sources.list.d/socktop.list
# Update and install
sudo apt update
sudo apt install socktop socktop-agent</code></pre>
<h2>Manual Installation</h2>
<p>Download packages from <a href="pool/main/">pool/main/</a></p>
<h2>Supported Architectures</h2>
<ul>
<li>amd64 (x86_64)</li>
<li>arm64 (aarch64)</li>
<li>armhf (32-bit ARM)</li>
</ul>
<h2>Source Code</h2>
<p>Visit the <a href="https://github.com/jasonwitty/socktop">GitHub repository</a></p>
</body>
</html>
Binary file not shown.
Binary file not shown.
+274
View File
@@ -0,0 +1,274 @@
# Debian Packaging for socktop
This document describes how to build and use Debian packages for socktop and socktop_agent.
## Prerequisites
Install `cargo-deb`:
```bash
cargo install cargo-deb
```
## Building Packages Locally
### Build for your current architecture (x86_64)
```bash
# Build socktop TUI client
cargo deb --package socktop
# Build socktop_agent daemon
cargo deb --package socktop_agent
```
The `.deb` files will be created in `target/debian/`.
### Cross-compile for ARM64 (Raspberry Pi, etc.)
First, install cross-compilation tools:
```bash
sudo apt-get update
sudo apt-get install gcc-aarch64-linux-gnu libc6-dev-arm64-cross
```
Add the ARM64 target:
```bash
rustup target add aarch64-unknown-linux-gnu
```
Configure the linker by creating `.cargo/config.toml`:
```toml
[target.aarch64-unknown-linux-gnu]
linker = "aarch64-linux-gnu-gcc"
```
Build the packages:
```bash
# Build for ARM64
cargo deb --package socktop --target aarch64-unknown-linux-gnu
cargo deb --package socktop_agent --target aarch64-unknown-linux-gnu
```
## Installing Packages
### Install socktop TUI client
```bash
sudo dpkg -i socktop_*.deb
```
### Install socktop_agent daemon
```bash
sudo dpkg -i socktop_agent_*.deb
```
The agent package will:
- Create a `socktop` system user and group
- Install the binary to `/usr/bin/socktop_agent`
- Install a systemd service file (disabled by default)
- Create `/var/lib/socktop` for state files
### Enable and start the agent service
```bash
# Enable to start on boot
sudo systemctl enable socktop-agent
# Start the service
sudo systemctl start socktop-agent
# Check status
sudo systemctl status socktop-agent
```
### Configure the agent
Edit the systemd service to customize settings:
```bash
sudo systemctl edit socktop-agent
```
Add configuration in the override section:
```ini
[Service]
Environment=SOCKTOP_PORT=8080
Environment=SOCKTOP_TOKEN=your-secret-token
Environment=RUST_LOG=info
```
Then restart:
```bash
sudo systemctl restart socktop-agent
```
## GitHub Actions
The project includes a GitHub Actions workflow (`.github/workflows/build-deb.yml`) that automatically builds `.deb` packages for both x86_64 and ARM64 architectures on every push to master or when tags are created.
### Downloading pre-built packages
1. Go to the [Actions tab](https://github.com/jasonwitty/socktop/actions)
2. Click on the latest "Build Debian Packages" workflow run
3. Download the artifacts:
- `debian-packages-amd64` - x86_64 packages
- `debian-packages-arm64` - ARM64 packages
- `all-debian-packages` - All packages combined
- `checksums` - SHA256 checksums
### Release packages
When you create a git tag starting with `v` (e.g., `v1.50.0`), the workflow will automatically create a GitHub Release with all `.deb` packages attached.
```bash
git tag v1.50.0
git push origin v1.50.0
```
## Package Details
### socktop package
- **Binary**: `/usr/bin/socktop`
- **Documentation**: `/usr/share/doc/socktop/`
- **Size**: ~5-8 MB (depends on architecture)
### socktop_agent package
- **Binary**: `/usr/bin/socktop_agent`
- **Service**: `socktop-agent.service`
- **User/Group**: `socktop`
- **State directory**: `/var/lib/socktop`
- **Config directory**: `/etc/socktop` (created but empty by default)
- **Documentation**: `/usr/share/doc/socktop_agent/`
- **Size**: ~5-8 MB (depends on architecture)
## Uninstalling
```bash
# Remove packages but keep configuration
sudo apt remove socktop socktop_agent
# Remove packages and all configuration (purge)
sudo apt purge socktop socktop_agent
```
When purging `socktop_agent`, the following are removed:
- The `socktop` user and group
- `/var/lib/socktop` directory
- Empty `/etc/socktop` directory (if empty)
## Verifying Packages
Check package contents:
```bash
dpkg -c socktop_*.deb
dpkg -c socktop_agent_*.deb
```
Check package information:
```bash
dpkg -I socktop_*.deb
dpkg -I socktop_agent_*.deb
```
After installation, verify files:
```bash
dpkg -L socktop
dpkg -L socktop-agent
```
## Troubleshooting
### Service fails to start
Check logs:
```bash
sudo journalctl -u socktop-agent -f
```
Verify the socktop user exists:
```bash
id socktop
```
### Permission issues
Ensure the state directory has correct permissions:
```bash
sudo chown -R socktop:socktop /var/lib/socktop
sudo chmod 755 /var/lib/socktop
```
### Missing dependencies
If installation fails due to missing dependencies:
```bash
sudo apt --fix-broken install
```
## Creating a Local APT Repository (Advanced)
To create your own APT repository for easy installation:
1. Install required tools:
```bash
sudo apt install dpkg-dev
```
2. Create repository structure:
```bash
mkdir -p ~/socktop-repo/pool/main
cp *.deb ~/socktop-repo/pool/main/
```
3. Generate package index:
```bash
cd ~/socktop-repo
dpkg-scanpackages pool/main /dev/null | gzip -9c > pool/main/Packages.gz
```
4. Serve via HTTP (for testing):
```bash
cd ~/socktop-repo
python3 -m http.server 8000
```
5. Add to sources on client machines:
```bash
echo "deb [trusted=yes] http://your-server:8000 pool/main/" | \
sudo tee /etc/apt/sources.list.d/socktop.list
sudo apt update
sudo apt install socktop socktop-agent
```
## Contributing
When adding new features that affect packaging:
1. Update `Cargo.toml` metadata in the `[package.metadata.deb]` section
2. Add new assets to the `assets` array if needed
3. Update maintainer scripts in `socktop_agent/debian/` if needed
4. Test package building locally before committing
5. Update this documentation
## References
- [cargo-deb documentation](https://github.com/kornelski/cargo-deb)
- [Debian Policy Manual](https://www.debian.org/doc/debian-policy/)
- [systemd service files](https://www.freedesktop.org/software/systemd/man/systemd.service.html)
+13 -10
View File
@@ -2,6 +2,8 @@
This guide explains how to cross-compile the socktop_agent on various host systems and deploy it to a Raspberry Pi. Cross-compiling is particularly useful for older or resource-constrained Pi models where native compilation might be slow.
**Note:** GPU monitoring support is not available on ARMv7 (32-bit) and RISC-V architectures due to library limitations. When building for these platforms, the `--no-default-features` flag must be used to disable GPU support.
## Cross-Compilation Host Setup
Choose your host operating system:
@@ -23,8 +25,9 @@ sudo apt update
sudo apt install gcc-aarch64-linux-gnu libc6-dev-arm64-cross libdrm-dev:arm64
# For 32-bit Raspberry Pi (armv7)
# Note: GPU support not available on armv7
sudo apt update
sudo apt install gcc-arm-linux-gnueabihf libc6-dev-armhf-cross libdrm-dev:armhf
sudo apt install gcc-arm-linux-gnueabihf libc6-dev-armhf-cross
```
### Setup Rust Cross-Compilation Targets
@@ -65,9 +68,8 @@ sudo pacman -S aarch64-linux-gnu-gcc
yay -S aarch64-linux-gnu-libdrm
# For 32-bit Raspberry Pi (armv7)
# Note: GPU support not available on armv7
sudo pacman -S arm-linux-gnueabihf-gcc
# Install libdrm for armv7 using an AUR helper
yay -S arm-linux-gnueabihf-libdrm
```
### Setup Rust Cross-Compilation Targets
@@ -114,8 +116,8 @@ cd path/to/socktop
# For 64-bit Raspberry Pi
docker run --rm -it -v "$(pwd)":/home/rust/src messense/rust-musl-cross:aarch64-musl cargo build --release --target aarch64-unknown-linux-musl -p socktop_agent
# For 32-bit Raspberry Pi
docker run --rm -it -v "$(pwd)":/home/rust/src messense/rust-musl-cross:armv7-musleabihf cargo build --release --target armv7-unknown-linux-musleabihf -p socktop_agent
# For 32-bit Raspberry Pi (without GPU support)
docker run --rm -it -v "$(pwd)":/home/rust/src messense/rust-musl-cross:armv7-musleabihf cargo build --release --target armv7-unknown-linux-musleabihf -p socktop_agent --no-default-features
```
The compiled binaries will be available in your local target directory.
@@ -133,11 +135,11 @@ The recommended approach for Windows is to use Windows Subsystem for Linux (WSL2
After setting up your environment, build the socktop_agent for your target Raspberry Pi:
```bash
# For 64-bit Raspberry Pi
# For 64-bit Raspberry Pi (with GPU support)
cargo build --release --target aarch64-unknown-linux-gnu -p socktop_agent
# For 32-bit Raspberry Pi
cargo build --release --target armv7-unknown-linux-gnueabihf -p socktop_agent
# For 32-bit Raspberry Pi (without GPU support)
cargo build --release --target armv7-unknown-linux-gnueabihf -p socktop_agent --no-default-features
```
## Transfer the Binary to Your Raspberry Pi
@@ -161,11 +163,12 @@ SSH into your Raspberry Pi and install the required dependencies:
```bash
ssh pi@raspberry-pi-ip
# For Raspberry Pi OS (Debian-based)
# For Raspberry Pi OS (Debian-based) - 64-bit only
# (32-bit armv7 builds don't require these)
sudo apt update
sudo apt install libdrm-dev libdrm-amdgpu1
# For Arch Linux ARM
# For Arch Linux ARM - 64-bit only
sudo pacman -Syu
sudo pacman -S libdrm
```
+26
View File
@@ -0,0 +1,26 @@
#!/usr/bin/env bash
set -euo pipefail
# Sync this repo to the 'gitea' remote as a mirror.
# - Mirrors ALL refs (branches, tags) and prunes removed ones.
# - This makes the Gitea repo match GitHub exactly.
if ! git rev-parse --is-inside-work-tree >/dev/null 2>&1; then
echo "Error: not inside a git repo" >&2
exit 1
fi
if ! git remote get-url gitea >/dev/null 2>&1; then
echo "Missing 'gitea' remote. Add it with:" >&2
echo " git remote add gitea https://gt.wittyoneoff.com/jason/socktop.git" >&2
exit 1
fi
echo "Fetching from origin (pruning)..."
git fetch origin --prune --tags
echo "Pushing mirror to gitea..."
git push gitea --mirror
echo "Done: Gitea should now match origin (GitHub)."
+19
View File
@@ -6,6 +6,8 @@ description = "Remote system monitor over WebSocket, TUI like top"
edition = "2024"
license = "MIT"
readme = "README.md"
homepage = "https://github.com/jasonwitty/socktop"
repository = "https://github.com/jasonwitty/socktop"
[dependencies]
# socktop connector for agent communication
@@ -25,3 +27,20 @@ sysinfo = { workspace = true }
[dev-dependencies]
assert_cmd = "2.0"
tempfile = "3"
[package.metadata.deb]
maintainer = "Jason Witty <jasonpwitty+socktop@proton.me>"
copyright = "2024, Jason Witty <jasonpwitty+socktop@proton.me>"
license-file = ["../LICENSE", "4"]
extended-description = """\
socktop is a remote system monitor with a rich terminal user interface (TUI) \
that connects to remote hosts running the socktop_agent over WebSocket. \
It provides real-time monitoring of CPU, memory, processes, and more with \
an interface similar to the traditional 'top' command."""
depends = "$auto"
section = "admin"
priority = "optional"
assets = [
["target/release/socktop", "usr/bin/", "755"],
["../README.md", "usr/share/doc/socktop/", "644"],
]
+211 -125
View File
@@ -29,11 +29,15 @@ use crate::ui::cpu::{
};
use crate::ui::modal::{ModalAction, ModalManager, ModalType};
use crate::ui::processes::{
ProcSortBy, ProcessKeyParams, get_filtered_sorted_indices, processes_handle_key_with_selection,
ProcSortBy, ProcessKeyParams, processes_handle_key_with_selection,
processes_handle_mouse_with_selection,
};
use crate::ui::{
disks::draw_disks, gpu::draw_gpu, header::draw_header, mem::draw_mem, net::draw_net_spark,
disks::draw_disks,
gpu::draw_gpu,
header::{build_header_intervals, build_header_title, draw_header},
mem::draw_mem,
net::draw_net_spark,
swap::draw_swap,
};
@@ -46,6 +50,15 @@ use socktop_connector::{
const MIN_METRICS_INTERVAL_MS: u64 = 100;
const MIN_PROCESSES_INTERVAL_MS: u64 = 200;
/// Drop duplicate-name entries from a disks payload (the agent occasionally
/// reports a partition twice). Done once when fresh disk data arrives so the
/// per-frame draw path doesn't have to rebuild a HashSet.
fn dedup_disks(disks: &mut Vec<socktop_connector::DiskInfo>) {
let mut seen: std::collections::HashSet<String> =
std::collections::HashSet::with_capacity(disks.len());
disks.retain(|d| seen.insert(d.name.clone()));
}
#[derive(Debug, Clone, PartialEq)]
pub enum ConnectionState {
Connected,
@@ -57,8 +70,9 @@ pub struct App {
// Latest metrics + histories
last_metrics: Option<Metrics>,
// CPU avg history (0..100)
// CPU avg history (0..100) with a running sum so draw avoids a 600-element fold per frame
cpu_hist: VecDeque<u64>,
cpu_hist_sum: u64,
// Per-core history (0..100)
per_core_hist: PerCoreHistory,
@@ -89,6 +103,17 @@ pub struct App {
pub process_search_active: bool,
pub process_search_query: String,
// Cached filtered + sorted process indices. Refreshed lazily when any of
// (metrics, sort order, search query) changes — input handlers, the draw
// path, and auto-scroll all read from this slice so we avoid rebuilding
// an indices Vec on every event.
procs_filtered: Vec<usize>,
procs_filter_dirty: bool,
// Pre-formatted process-row strings, rebuilt once per procs poll. Indexed
// parallel to `last_metrics.top_processes`.
procs_row_cache: Vec<crate::ui::processes::CachedRow>,
procs_row_peak_cpu: f32,
last_procs_poll: Instant,
last_disks_poll: Instant,
procs_interval: Duration,
@@ -99,6 +124,7 @@ pub struct App {
pub process_details: Option<socktop_connector::ProcessMetricsResponse>,
pub journal_entries: Option<socktop_connector::JournalResponse>,
pub process_cpu_history: VecDeque<f32>, // CPU history for sparkline (last 60 samples)
pub process_cpu_history_sum: f32, // running sum of process_cpu_history
pub process_mem_history: VecDeque<u64>, // Memory usage history in bytes (last 60 samples)
pub process_io_read_history: VecDeque<u64>, // Disk read DELTA history in bytes (last 60 samples)
pub process_io_write_history: VecDeque<u64>, // Disk write DELTA history in bytes (last 60 samples)
@@ -119,6 +145,17 @@ pub struct App {
pub is_tls: bool,
pub has_token: bool,
// Cached title strings — only rebuilt when source values change so the
// diff renderer can suppress redraws on idle frames.
header_title: String,
header_title_key: (String, bool, bool),
header_intervals_text: String,
header_intervals_key: (u128, u128),
net_dl_title: String,
net_dl_key: (u64, u64),
net_ul_title: String,
net_ul_key: (u64, u64),
// Modal system
pub modal_manager: crate::ui::modal::ModalManager,
@@ -136,6 +173,7 @@ impl App {
Self {
last_metrics: None,
cpu_hist: VecDeque::with_capacity(600),
cpu_hist_sum: 0,
per_core_hist: PerCoreHistory::new(60),
last_net_totals: None,
rx_hist: VecDeque::with_capacity(600),
@@ -154,6 +192,10 @@ impl App {
prev_selected_process_pid: None,
process_search_active: false,
process_search_query: String::new(),
procs_filtered: Vec::new(),
procs_filter_dirty: true,
procs_row_cache: Vec::new(),
procs_row_peak_cpu: 0.0,
last_procs_poll: Instant::now()
.checked_sub(Duration::from_secs(2))
.unwrap_or_else(Instant::now), // trigger immediately on first loop
@@ -166,6 +208,7 @@ impl App {
process_details: None,
journal_entries: None,
process_cpu_history: VecDeque::with_capacity(600),
process_cpu_history_sum: 0.0,
process_mem_history: VecDeque::with_capacity(600),
process_io_read_history: VecDeque::with_capacity(600),
process_io_write_history: VecDeque::with_capacity(600),
@@ -186,6 +229,14 @@ impl App {
verify_hostname: false,
is_tls: false,
has_token: false,
header_title: String::new(),
header_title_key: (String::new(), false, false),
header_intervals_text: String::new(),
header_intervals_key: (u128::MAX, u128::MAX),
net_dl_title: String::new(),
net_dl_key: (u64::MAX, u64::MAX),
net_ul_title: String::new(),
net_ul_key: (u64::MAX, u64::MAX),
modal_manager: ModalManager::new(),
connection_state: ConnectionState::Disconnected,
last_connection_attempt: Instant::now(),
@@ -465,7 +516,10 @@ impl App {
_url: &str,
_tls_ca: Option<&str>,
_verify_hostname: bool,
) -> Result<(), Box<dyn std::error::Error>> {
) -> Result<(), Box<dyn std::error::Error>>
where
<B as ratatui::backend::Backend>::Error: 'static,
{
loop {
// Handle input for modal
while event::poll(Duration::from_millis(10))? {
@@ -572,7 +626,10 @@ impl App {
&mut self,
terminal: &mut Terminal<B>,
mut ws: SocktopConnector,
) -> Result<(), Box<dyn std::error::Error>> {
) -> Result<(), Box<dyn std::error::Error>>
where
<B as ratatui::backend::Backend>::Error: 'static,
{
loop {
// Main event loop
let result = self.run_event_loop_iteration(terminal, &mut ws).await;
@@ -592,7 +649,10 @@ impl App {
&mut self,
terminal: &mut Terminal<B>,
ws: &mut SocktopConnector,
) -> Result<(), Box<dyn std::error::Error>> {
) -> Result<(), Box<dyn std::error::Error>>
where
<B as ratatui::backend::Backend>::Error: 'static,
{
loop {
// Input (non-blocking)
while event::poll(Duration::from_millis(10))? {
@@ -665,6 +725,7 @@ impl App {
// Exit search mode
self.process_search_active = false;
self.process_search_query.clear();
self.invalidate_procs_filter();
continue;
}
KeyCode::Enter => {
@@ -672,27 +733,24 @@ impl App {
self.process_search_active = false;
// Auto-select first filtered result
if let Some(m) = self.last_metrics.as_ref() {
let idxs = get_filtered_sorted_indices(
m,
&self.process_search_query,
self.procs_sort_by,
);
if !idxs.is_empty() {
let first_idx = idxs[0];
self.selected_process_index = Some(first_idx);
self.selected_process_pid =
Some(m.top_processes[first_idx].pid);
}
let first = self.procs_filter().first().copied();
if let (Some(first_idx), Some(m)) =
(first, self.last_metrics.as_ref())
{
self.selected_process_index = Some(first_idx);
self.selected_process_pid =
Some(m.top_processes[first_idx].pid);
}
continue;
}
KeyCode::Backspace => {
self.process_search_query.pop();
self.invalidate_procs_filter();
continue;
}
KeyCode::Char(c) => {
self.process_search_query.push(c);
self.invalidate_procs_filter();
continue;
}
KeyCode::Up | KeyCode::Down => {
@@ -728,6 +786,7 @@ impl App {
self.process_search_query.clear();
self.selected_process_pid = None;
self.selected_process_index = None;
self.invalidate_procs_filter();
continue;
}
@@ -760,6 +819,10 @@ impl App {
.split(rows[1]);
let content = per_core_content_area(top[1]);
// Refresh the filtered+sorted index cache once before we
// borrow individual fields of `self`.
let _ = self.procs_filter();
// First try process selection (only handles arrows if a process is selected)
let process_handled = if self.last_procs_area.is_some() {
processes_handle_key_with_selection(ProcessKeyParams {
@@ -767,8 +830,7 @@ impl App {
selected_process_index: &mut self.selected_process_index,
key: k,
metrics: self.last_metrics.as_ref(),
sort_by: self.procs_sort_by,
search_query: &self.process_search_query,
filtered_indices: &self.procs_filtered,
})
} else {
false
@@ -786,14 +848,9 @@ impl App {
// Auto-scroll to keep selected process visible
if let (Some(selected_idx), Some(p_area)) =
(self.selected_process_index, self.last_procs_area)
&& let Some(m) = self.last_metrics.as_ref()
&& self.last_metrics.is_some()
{
// Get filtered and sorted indices (same as display)
let idxs = get_filtered_sorted_indices(
m,
&self.process_search_query,
self.procs_sort_by,
);
let idxs = &self.procs_filtered;
// Find the display position of the selected process in filtered list
if let Some(display_pos) =
@@ -903,11 +960,17 @@ impl App {
content.height as usize,
);
// Refresh filter cache before partial borrows of self.
let _ = self.procs_filter();
let search_box_visible =
self.process_search_active || !self.process_search_query.is_empty();
// Processes table: sort by column on header click and handle row selection
if let (Some(mm), Some(p_area)) =
if let (Some(_mm), Some(p_area)) =
(self.last_metrics.as_ref(), self.last_procs_area)
{
use crate::ui::processes::ProcessMouseParams;
let total_rows = self.procs_filtered.len();
if let Some(new_sort) =
processes_handle_mouse_with_selection(ProcessMouseParams {
scroll_offset: &mut self.procs_scroll_offset,
@@ -916,13 +979,14 @@ impl App {
drag: &mut self.procs_drag,
mouse: m,
area: p_area,
total_rows: mm.top_processes.len(),
total_rows,
metrics: self.last_metrics.as_ref(),
sort_by: self.procs_sort_by,
search_query: &self.process_search_query,
search_box_visible,
filtered_indices: &self.procs_filtered,
})
{
self.procs_sort_by = new_sort;
self.invalidate_procs_filter();
}
}
@@ -959,22 +1023,36 @@ impl App {
// Only poll processes every 2s
if self.last_procs_poll.elapsed() >= self.procs_interval {
let mut updated = false;
if let Ok(AgentResponse::Processes(procs)) =
ws.request(AgentRequest::Processes).await
&& let Some(mm) = self.last_metrics.as_mut()
{
mm.top_processes = procs.top_processes;
mm.process_count = Some(procs.process_count);
updated = true;
}
if updated {
self.invalidate_procs_filter();
// Rebuild the pre-formatted row cache for the next
// ~N frames. Done once per poll, not per frame.
if let Some(mm) = self.last_metrics.as_ref() {
self.procs_row_peak_cpu = crate::ui::processes::rebuild_row_cache(
mm,
&mut self.procs_row_cache,
);
}
}
self.last_procs_poll = Instant::now();
}
// Only poll disks every 5s
if self.last_disks_poll.elapsed() >= self.disks_interval {
if let Ok(AgentResponse::Disks(disks)) =
if let Ok(AgentResponse::Disks(mut disks)) =
ws.request(AgentRequest::Disks).await
&& let Some(mm) = self.last_metrics.as_mut()
{
dedup_disks(&mut disks);
mm.disks = disks;
}
self.last_disks_poll = Instant::now();
@@ -1000,7 +1078,14 @@ impl App {
Ok(Ok(AgentResponse::ProcessMetrics(details))) => {
// Update history for sparklines
let cpu_usage = details.process.cpu_usage;
push_capped(&mut self.process_cpu_history, cpu_usage, 600);
let evicted_cpu = push_capped(
&mut self.process_cpu_history,
cpu_usage,
600,
);
self.process_cpu_history_sum = self.process_cpu_history_sum
+ cpu_usage
- evicted_cpu.unwrap_or(0.0);
let mem_bytes = details.process.mem_bytes;
push_capped(&mut self.process_mem_history, mem_bytes, 600);
@@ -1105,11 +1190,37 @@ impl App {
Ok(())
}
/// Mark the filtered-process cache stale. Call this whenever
/// `procs_sort_by`, `process_search_query`, or the top_processes content
/// changes — the cache is rebuilt lazily on the next read.
pub fn invalidate_procs_filter(&mut self) {
self.procs_filter_dirty = true;
}
/// Lazily refresh and return the cached filtered+sorted process indices.
/// Empty slice when there are no metrics yet.
pub fn procs_filter(&mut self) -> &[usize] {
if self.procs_filter_dirty {
self.procs_filtered.clear();
if let Some(m) = self.last_metrics.as_ref() {
crate::ui::processes::fill_filtered_sorted_indices(
m,
&self.process_search_query,
self.procs_sort_by,
&mut self.procs_filtered,
);
}
self.procs_filter_dirty = false;
}
&self.procs_filtered
}
/// Clear process details when modal is closed or selection changes
pub fn clear_process_details(&mut self) {
self.process_details = None;
self.journal_entries = None;
self.process_cpu_history.clear();
self.process_cpu_history_sum = 0.0;
self.process_mem_history.clear();
self.process_io_read_history.clear();
self.process_io_write_history.clear();
@@ -1120,23 +1231,24 @@ impl App {
}
fn update_with_metrics(&mut self, mut m: Metrics) {
if let Some(prev) = &self.last_metrics {
// Preserve slower fields when the fast payload omits them
if let Some(prev) = self.last_metrics.as_mut() {
// Preserve slower fields when the fast payload omits them.
// prev is about to be dropped so we can move its Vecs instead of cloning.
if m.disks.is_empty() {
m.disks = prev.disks.clone();
m.disks = std::mem::take(&mut prev.disks);
}
if m.top_processes.is_empty() {
m.top_processes = prev.top_processes.clone();
m.top_processes = std::mem::take(&mut prev.top_processes);
}
// Preserve total processes count across fast updates
if m.process_count.is_none() {
m.process_count = prev.process_count;
}
}
// CPU avg history
// CPU avg history with running sum
let v = m.cpu_total.clamp(0.0, 100.0).round() as u64;
push_capped(&mut self.cpu_hist, v, 600);
let evicted = push_capped(&mut self.cpu_hist, v, 600);
self.cpu_hist_sum = self.cpu_hist_sum + v - evicted.unwrap_or(0);
// Per-core history (push current samples)
self.per_core_hist.ensure_cores(m.cpu_per_core.len());
@@ -1179,16 +1291,31 @@ impl App {
])
.split(area);
// Header
draw_header(
f,
rows[0],
self.last_metrics.as_ref(),
self.is_tls,
self.has_token,
self.metrics_interval,
self.procs_interval,
);
// Header — refresh cached strings only when their inputs change so the
// ratatui diff renderer can suppress repaints on idle frames.
{
let hostname = self.last_metrics.as_ref().map(|mm| mm.hostname.as_str());
let key = (
hostname.unwrap_or("").to_string(),
self.is_tls,
self.has_token,
);
if self.header_title_key != key {
self.header_title = build_header_title(hostname, self.is_tls, self.has_token);
self.header_title_key = key;
}
let intervals_key = (
self.metrics_interval.as_millis(),
self.procs_interval.as_millis(),
);
if self.header_intervals_key != intervals_key {
self.header_intervals_text =
build_header_intervals(intervals_key.0, intervals_key.1);
self.header_intervals_key = intervals_key;
}
}
draw_header(f, rows[0], &self.header_title, &self.header_intervals_text);
// Top row: left CPU avg, right Per-core (full top-right)
let top_lr = ratatui::layout::Layout::default()
@@ -1196,12 +1323,18 @@ impl App {
.constraints([Constraint::Percentage(66), Constraint::Percentage(34)])
.split(rows[1]);
draw_cpu_avg_graph(f, top_lr[0], &self.cpu_hist, self.last_metrics.as_ref());
draw_cpu_avg_graph(
f,
top_lr[0],
&mut self.cpu_hist,
self.cpu_hist_sum,
self.last_metrics.as_ref(),
);
draw_per_core_bars(
f,
top_lr[1],
self.last_metrics.as_ref(),
&self.per_core_hist,
&mut self.per_core_hist,
self.per_core_scroll,
);
@@ -1245,26 +1378,33 @@ impl App {
.split(bottom_lr[0]);
draw_disks(f, left_stack[0], self.last_metrics.as_ref());
// Net titles only change when the throughput or peak changes.
let rx_now = self.rx_hist.back().copied().unwrap_or(0);
let rx_key = (rx_now, self.rx_peak);
if self.net_dl_key != rx_key {
self.net_dl_title = format!("Download (KB/s) — now: {rx_now} | peak: {}", self.rx_peak);
self.net_dl_key = rx_key;
}
draw_net_spark(
f,
left_stack[1],
&format!(
"Download (KB/s) — now: {} | peak: {}",
self.rx_hist.back().copied().unwrap_or(0),
self.rx_peak
),
&self.rx_hist,
&self.net_dl_title,
&mut self.rx_hist,
ratatui::style::Color::Green,
);
let tx_now = self.tx_hist.back().copied().unwrap_or(0);
let tx_key = (tx_now, self.tx_peak);
if self.net_ul_key != tx_key {
self.net_ul_title = format!("Upload (KB/s) — now: {tx_now} | peak: {}", self.tx_peak);
self.net_ul_key = tx_key;
}
draw_net_spark(
f,
left_stack[2],
&format!(
"Upload (KB/s) — now: {} | peak: {}",
self.tx_hist.back().copied().unwrap_or(0),
self.tx_peak
),
&self.tx_hist,
&self.net_ul_title,
&mut self.tx_hist,
ratatui::style::Color::Blue,
);
@@ -1272,6 +1412,8 @@ impl App {
let procs_area = bottom_lr[1];
// Cache for input handlers
self.last_procs_area = Some(procs_area);
// Refresh the filter cache before partial borrows of self.
let _ = self.procs_filter();
crate::ui::processes::draw_top_processes(
f,
procs_area,
@@ -1283,6 +1425,9 @@ impl App {
selected_process_index: self.selected_process_index,
search_query: &self.process_search_query,
search_active: self.process_search_active,
filtered_indices: &self.procs_filtered,
cached_rows: &self.procs_row_cache,
peak_cpu: self.procs_row_peak_cpu,
},
);
@@ -1296,6 +1441,7 @@ impl App {
journal: self.journal_entries.as_ref(),
history: ProcessHistoryData {
cpu: &self.process_cpu_history,
cpu_sum: self.process_cpu_history_sum,
mem: &self.process_mem_history,
io_read: &self.process_io_read_history,
io_write: &self.process_io_write_history,
@@ -1310,66 +1456,6 @@ impl App {
impl Default for App {
fn default() -> Self {
Self {
last_metrics: None,
cpu_hist: VecDeque::with_capacity(600),
per_core_hist: PerCoreHistory::new(60),
last_net_totals: None,
rx_hist: VecDeque::with_capacity(600),
tx_hist: VecDeque::with_capacity(600),
rx_peak: 0,
tx_peak: 0,
should_quit: false,
per_core_scroll: 0,
per_core_drag: None,
procs_scroll_offset: 0,
procs_drag: None,
procs_sort_by: ProcSortBy::CpuDesc,
last_procs_area: None,
selected_process_pid: None,
selected_process_index: None,
prev_selected_process_pid: None,
process_search_active: false,
process_search_query: String::new(),
last_procs_poll: Instant::now()
.checked_sub(Duration::from_secs(2))
.unwrap_or_else(Instant::now), // trigger immediately on first loop
last_disks_poll: Instant::now()
.checked_sub(Duration::from_secs(5))
.unwrap_or_else(Instant::now),
procs_interval: Duration::from_secs(2),
disks_interval: Duration::from_secs(5),
metrics_interval: Duration::from_millis(500),
process_details: None,
journal_entries: None,
process_cpu_history: VecDeque::with_capacity(600),
process_mem_history: VecDeque::with_capacity(600),
process_io_read_history: VecDeque::with_capacity(600),
process_io_write_history: VecDeque::with_capacity(600),
last_io_read_bytes: None,
last_io_write_bytes: None,
max_process_mem_bytes: 0,
process_details_unsupported: false,
last_process_details_poll: Instant::now()
.checked_sub(Duration::from_secs(10))
.unwrap_or_else(Instant::now),
last_journal_poll: Instant::now()
.checked_sub(Duration::from_secs(10))
.unwrap_or_else(Instant::now),
process_details_interval: Duration::from_millis(500),
journal_interval: Duration::from_secs(5),
ws_url: String::new(),
tls_ca: None,
verify_hostname: false,
is_tls: false,
has_token: false,
modal_manager: ModalManager::new(),
connection_state: ConnectionState::Disconnected,
last_connection_attempt: Instant::now(),
original_disconnect_time: None,
connection_retry_count: 0,
last_auto_retry: None,
replacement_connection: None,
}
Self::new()
}
}
+16 -7
View File
@@ -2,16 +2,25 @@
use std::collections::VecDeque;
pub fn push_capped<T>(dq: &mut VecDeque<T>, v: T, cap: usize) {
if dq.len() == cap {
dq.pop_front();
}
/// Push a value into a capped deque. Returns the evicted front element if any.
/// Callers maintaining a running sum can use this to update the sum without
/// re-iterating the whole deque.
pub fn push_capped<T>(dq: &mut VecDeque<T>, v: T, cap: usize) -> Option<T> {
let evicted = if dq.len() == cap {
dq.pop_front()
} else {
None
};
dq.push_back(v);
evicted
}
// Keeps a history deque per core with a fixed capacity
// Keeps a history deque per core with a fixed capacity.
// Storage is u64 so sparkline rendering can hand the slice directly to
// ratatui's `Sparkline::data` (which takes `&[u64]`) without per-frame
// allocation or widening conversion.
pub struct PerCoreHistory {
pub deques: Vec<VecDeque<u16>>,
pub deques: Vec<VecDeque<u64>>,
cap: usize,
}
@@ -35,7 +44,7 @@ impl PerCoreHistory {
pub fn push_samples(&mut self, samples: &[f32]) {
self.ensure_cores(samples.len());
for (i, v) in samples.iter().enumerate() {
let val = v.clamp(0.0, 100.0).round() as u16;
let val = v.clamp(0.0, 100.0).round() as u64;
push_capped(&mut self.deques[i], val, self.cap);
}
}
+158 -67
View File
@@ -7,7 +7,9 @@ use ratatui::style::{Color, Style};
use ratatui::{
layout::{Constraint, Direction, Layout, Rect},
text::{Line, Span},
widgets::{Block, Borders, Paragraph, Sparkline},
widgets::{
Block, Borders, Paragraph, Scrollbar, ScrollbarOrientation, ScrollbarState, Sparkline,
},
};
use crate::history::PerCoreHistory;
@@ -133,11 +135,9 @@ pub fn per_core_handle_scrollbar_mouse(
}
let thumb_len = (track * view).div_ceil(total).max(1).min(track);
let top_for_offset = |off: usize| -> usize {
if max_off == 0 {
0
} else {
((track - thumb_len) * off + max_off / 2) / max_off
}
((track - thumb_len) * off + max_off / 2)
.checked_div(max_off)
.unwrap_or(0)
};
let thumb_top = top_for_offset(offset);
@@ -190,11 +190,9 @@ pub fn per_core_handle_scrollbar_mouse(
// Inverse mapping top -> offset
if track > thumb_len {
let denom = track - thumb_len;
offset = if max_off == 0 {
0
} else {
(new_top * max_off + denom / 2) / denom
};
offset = (new_top * max_off + denom / 2)
.checked_div(denom)
.unwrap_or(0);
} else {
offset = 0;
}
@@ -234,18 +232,20 @@ pub fn per_core_clamp(scroll_offset: &mut usize, total_rows: usize, viewport_row
}
/// Draws the CPU average sparkline graph.
///
/// `hist_sum` is the running sum of `hist` maintained by the caller so we don't
/// fold the (up to 600-element) deque on every frame.
pub fn draw_cpu_avg_graph(
f: &mut ratatui::Frame<'_>,
area: Rect,
hist: &std::collections::VecDeque<u64>,
hist: &mut std::collections::VecDeque<u64>,
hist_sum: u64,
m: Option<&Metrics>,
) {
// Calculate average CPU over the monitoring period
let avg_cpu = if !hist.is_empty() {
let sum: u64 = hist.iter().sum();
sum as f64 / hist.len() as f64
} else {
let avg_cpu = if hist.is_empty() {
0.0
} else {
hist_sum as f64 / hist.len() as f64
};
let title = if let Some(mm) = m {
@@ -272,14 +272,16 @@ pub fn draw_cpu_avg_graph(
String::new()
};
// Hand a slice directly to Sparkline. `make_contiguous` is amortized cheap
// for our usage pattern (cap'd 600-element ring updated at 2 Hz) and lets
// us skip the per-frame Vec allocation .collect() used to do.
let max_points = area.width.saturating_sub(2) as usize;
let start = hist.len().saturating_sub(max_points);
let data: Vec<u64> = hist.iter().skip(start).cloned().collect();
let slice = &hist.make_contiguous()[start..];
// Render the sparkline with title on left
let spark = Sparkline::default()
.block(Block::default().borders(Borders::ALL).title(title))
.data(&data)
.data(slice)
.max(100)
.style(Style::default().fg(Color::Cyan));
f.render_widget(spark, area);
@@ -302,7 +304,7 @@ pub fn draw_per_core_bars(
f: &mut ratatui::Frame<'_>,
area: Rect,
m: Option<&Metrics>,
per_core_hist: &PerCoreHistory,
per_core_hist: &mut PerCoreHistory,
scroll_offset: usize,
) {
f.render_widget(
@@ -347,7 +349,7 @@ pub fn draw_per_core_bars(
let rect = vchunks[i];
let hchunks = Layout::default()
.direction(Direction::Horizontal)
.constraints([Constraint::Min(6), Constraint::Length(12)])
.constraints([Constraint::Min(6), Constraint::Length(13)])
.split(rect);
let curr = mm.cpu_per_core[idx].clamp(0.0, 100.0);
@@ -358,12 +360,17 @@ pub fn draw_per_core_bars(
.map(|v| v as f32)
.unwrap_or(curr);
// Trend indicator. Various Unicode glyphs we tried for the "flat"
// trend (╌, ·) substituted as a hyphen on terminals with narrow font
// coverage; combined with the next column being `100.0` they read as
// `cpu0 -100.0%`, a nonsensical negative percent. Use a literal space
// for the flat case — no character, no fallback, no confusion.
let trend = if curr > older + 0.2 {
""
} else if curr + 0.2 < older {
""
} else {
""
" "
};
let fg = match curr {
@@ -372,24 +379,24 @@ pub fn draw_per_core_bars(
_ => Color::Red,
};
let hist: Vec<u64> = per_core_hist
.deques
.get(idx)
.map(|d| {
let max_points = hchunks[0].width as usize;
let start = d.len().saturating_sub(max_points);
d.iter().skip(start).map(|&v| v as u64).collect()
})
.unwrap_or_default();
// Borrow the per-core deque mutably so we can hand a contiguous slice
// to Sparkline without allocating a fresh Vec each frame.
if let Some(d) = per_core_hist.deques.get_mut(idx) {
let max_points = hchunks[0].width as usize;
let start = d.len().saturating_sub(max_points);
let slice = &d.make_contiguous()[start..];
let spark = Sparkline::default()
.data(slice)
.max(100)
.style(Style::default().fg(fg));
f.render_widget(spark, hchunks[0]);
}
let spark = Sparkline::default()
.data(&hist)
.max(100)
.style(Style::default().fg(fg));
f.render_widget(spark, hchunks[0]);
let label = format!("cpu{idx:<2}{trend}{curr:>5.1}%");
// Hard space between the trend mark and the number — even if the
// arrow glyphs (↑/↓) fall back to ASCII on a terminal that lacks
// them, this space prevents the trend mark from visually joining
// `100.0` to look like a negative value.
let label = format!("cpu{idx:<2}{trend} {curr:>5.1}%");
let line = Line::from(Span::styled(
label,
Style::default().fg(fg).add_modifier(Modifier::BOLD),
@@ -397,38 +404,122 @@ pub fn draw_per_core_bars(
f.render_widget(Paragraph::new(line).right_aligned(), hchunks[1]);
}
// Custom 1-col scrollbar with arrows, track, and exact mapping
// 1-col scrollbar (ratatui built-in widget). Skips drawing when the
// content fits in the viewport, matching the previous behaviour.
let scroll_area = Rect {
x: inner.x + inner.width.saturating_sub(1),
y: inner.y,
width: 1,
height: inner.height,
};
if scroll_area.height >= 3 {
let track = (scroll_area.height - 2) as usize;
let total = total_rows.max(1);
let view = viewport_rows.clamp(1, total);
let max_off = total.saturating_sub(view);
let thumb_len = (track * view).div_ceil(total).max(1).min(track);
let thumb_top = if max_off == 0 {
0
} else {
((track - thumb_len) * offset + max_off / 2) / max_off
};
// Build lines: top arrow, track (with thumb), bottom arrow
let mut lines: Vec<Line> = Vec::with_capacity(scroll_area.height as usize);
lines.push(Line::from(Span::styled("", Style::default().fg(SB_ARROW))));
for i in 0..track {
if i >= thumb_top && i < thumb_top + thumb_len {
lines.push(Line::from(Span::styled("", Style::default().fg(SB_THUMB))));
} else {
lines.push(Line::from(Span::styled("", Style::default().fg(SB_TRACK))));
}
}
lines.push(Line::from(Span::styled("", Style::default().fg(SB_ARROW))));
f.render_widget(Paragraph::new(lines), scroll_area);
let max_off = total_rows.saturating_sub(viewport_rows);
if scroll_area.height >= 3 && max_off > 0 {
let scrollbar = Scrollbar::new(ScrollbarOrientation::VerticalRight)
.begin_symbol(Some(""))
.end_symbol(Some(""))
.thumb_symbol("")
.track_symbol(Some(""))
.thumb_style(Style::default().fg(SB_THUMB))
.track_style(Style::default().fg(SB_TRACK))
.begin_style(Style::default().fg(SB_ARROW))
.end_style(Style::default().fg(SB_ARROW));
let mut state = ScrollbarState::new(max_off).position(offset);
f.render_stateful_widget(scrollbar, scroll_area, &mut state);
}
}
#[cfg(test)]
mod render_tests {
use super::*;
use ratatui::Terminal;
use ratatui::backend::TestBackend;
use socktop_connector::Metrics;
fn fake_metrics(cores: Vec<f32>) -> Metrics {
Metrics {
cpu_total: 0.0,
cpu_per_core: cores,
mem_total: 1024,
mem_used: 0,
swap_total: 0,
swap_used: 0,
hostname: "t".into(),
cpu_temp_c: None,
disks: vec![],
networks: vec![],
top_processes: vec![],
gpus: None,
process_count: Some(0),
}
}
fn dump(terminal: &Terminal<TestBackend>) -> String {
let buf = terminal.backend().buffer();
let mut out = String::new();
for y in 0..buf.area().height {
for x in 0..buf.area().width {
out.push_str(buf[(x, y)].symbol());
}
out.push('\n');
}
out
}
/// Regression: the "flat" trend glyph used to be `╌` (U+254C), then `·`
/// (U+00B7) — both substituted as a hyphen on terminals with narrow font
/// coverage. When a core sat at exactly 100% the label rendered as
/// `cpu3 -100.0%` (no space between trend and digits). Now we use a
/// literal space for the flat case AND insert a hard space between every
/// trend mark and the number, so no glyph substitution can produce a
/// "-100" substring. We assert that across flat AND transitioning cores.
#[test]
fn percore_label_never_renders_as_negative() {
let m = fake_metrics(vec![100.0, 100.0, 100.0, 100.0]);
let mut hist = PerCoreHistory::new(60);
hist.ensure_cores(4);
// First sample: history is empty, no trend on first frame.
hist.push_samples(&m.cpu_per_core);
// Second sample: identical values → flat trend (the user's complaint).
hist.push_samples(&m.cpu_per_core);
let backend = TestBackend::new(120, 8);
let mut terminal = Terminal::new(backend).unwrap();
terminal
.draw(|f| {
draw_per_core_bars(f, Rect::new(0, 0, 120, 8), Some(&m), &mut hist, 0);
})
.unwrap();
let out = dump(&terminal);
eprintln!("---flat 100% render---\n{out}");
assert!(!out.contains("-100"), "found '-100' in flat-trend render");
// Decreasing trend at saturation: hist was high, current drops a bit.
let mut hist2 = PerCoreHistory::new(60);
hist2.ensure_cores(4);
for _ in 0..25 {
hist2.push_samples(&[100.0, 100.0, 100.0, 100.0]);
}
let m2 = fake_metrics(vec![100.0, 100.0, 100.0, 80.0]);
hist2.push_samples(&m2.cpu_per_core);
let backend = TestBackend::new(120, 8);
let mut terminal = Terminal::new(backend).unwrap();
terminal
.draw(|f| {
draw_per_core_bars(f, Rect::new(0, 0, 120, 8), Some(&m2), &mut hist2, 0);
})
.unwrap();
let out = dump(&terminal);
eprintln!("---decreasing render---\n{out}");
assert!(
!out.contains("-100"),
"found '-100' in decreasing-trend render"
);
assert!(
!out.contains("-80"),
"found '-80' in decreasing-trend render"
);
}
}
+5 -10
View File
@@ -24,16 +24,11 @@ pub fn draw_disks(f: &mut ratatui::Frame<'_>, area: Rect, m: Option<&Metrics>) {
return;
}
// Filter duplicates by keeping first occurrence of each unique name
let mut seen_names = std::collections::HashSet::new();
let unique_disks: Vec<_> = mm
.disks
.iter()
.filter(|d| seen_names.insert(d.name.clone()))
.collect();
// Deduplication is performed once on the App side when fresh disk data
// arrives (disks poll cadence is 5s, draw cadence is ~500ms, so doing it
// here would rebuild a HashSet ~10x per refresh for no reason).
let per_disk_h = 3u16;
let max_cards = (inner.height / per_disk_h).min(unique_disks.len() as u16) as usize;
let max_cards = (inner.height / per_disk_h).min(mm.disks.len() as u16) as usize;
let constraints: Vec<Constraint> = (0..max_cards)
.map(|_| Constraint::Length(per_disk_h))
@@ -44,7 +39,7 @@ pub fn draw_disks(f: &mut ratatui::Frame<'_>, area: Rect, m: Option<&Metrics>) {
.split(inner);
for (i, slot) in rows.iter().enumerate() {
let d = unique_disks[i];
let d = &mm.disks[i];
let used = d.total.saturating_sub(d.available);
let ratio = if d.total > 0 {
used as f64 / d.total as f64
+16 -27
View File
@@ -1,47 +1,36 @@
//! Top header with hostname and CPU temperature indicator.
use crate::types::Metrics;
use ratatui::{
layout::Rect,
text::{Line, Span},
widgets::{Block, Borders, Paragraph},
};
use std::time::Duration;
pub fn draw_header(
f: &mut ratatui::Frame<'_>,
area: Rect,
m: Option<&Metrics>,
is_tls: bool,
has_token: bool,
metrics_interval: Duration,
procs_interval: Duration,
) {
let base = if let Some(mm) = m {
format!("socktop — host: {}", mm.hostname)
} else {
"socktop — connecting...".into()
/// Build the header's left-side title from session state. Callers cache the
/// returned String and only rebuild it when one of the inputs changes.
pub fn build_header_title(hostname: Option<&str>, is_tls: bool, has_token: bool) -> String {
let base = match hostname {
Some(h) => format!("socktop — host: {h}"),
None => "socktop — connecting...".into(),
};
// TLS indicator: lock vs lock with cross (using ✗). Keep explicit label for clarity.
let tls_txt = if is_tls { "🔒 TLS" } else { "🔒✗ TLS" };
// Token indicator
let tok_txt = if has_token { "🔑 token" } else { "" };
let mut parts = vec![base, tls_txt.into()];
if !tok_txt.is_empty() {
parts.push(tok_txt.into());
if has_token {
parts.push("🔑 token".into());
}
parts.push("(a: about, h: help, q: quit)".into());
let title = parts.join(" | ");
parts.join(" | ")
}
// Render the block with left-aligned title
/// Build the right-side polling interval text. Callers cache this string.
pub fn build_header_intervals(metrics_ms: u128, procs_ms: u128) -> String {
format!("{metrics_ms}ms metrics | {procs_ms}ms procs")
}
pub fn draw_header(f: &mut ratatui::Frame<'_>, area: Rect, title: &str, intervals: &str) {
f.render_widget(Block::default().title(title).borders(Borders::BOTTOM), area);
// Render polling intervals on the right side
let mi = metrics_interval.as_millis();
let pi = procs_interval.as_millis();
let intervals = format!("{mi}ms metrics | {pi}ms procs");
let intervals_width = intervals.len() as u16;
if area.width > intervals_width + 2 {
let right_area = Rect {
x: area.x + area.width.saturating_sub(intervals_width + 1),
+6 -5
View File
@@ -41,11 +41,12 @@ impl ModalManager {
])
.split(area);
let block = Block::default()
.title(ICON_WARNING_TITLE)
.title_style(
Style::default()
.fg(MODAL_TITLE_FG)
.add_modifier(Modifier::BOLD),
.title(
Line::from(ICON_WARNING_TITLE).style(
Style::default()
.fg(MODAL_TITLE_FG)
.add_modifier(Modifier::BOLD),
),
)
.borders(Borders::ALL)
.border_style(Style::default().fg(MODAL_BORDER_FG))
+39 -26
View File
@@ -60,6 +60,7 @@ impl ModalManager {
main_chunks[0],
&details.process,
data.history.cpu,
data.history.cpu_sum,
);
// Middle Row: Memory/IO + Thread Table + Command Details (with process metadata)
@@ -419,7 +420,7 @@ impl ModalManager {
)
.header(header)
.block(block)
.highlight_style(Style::default());
.row_highlight_style(Style::default());
f.render_widget(table, area);
@@ -564,8 +565,13 @@ impl ModalManager {
return;
}
// Create a 2D grid to represent the plot
let mut plot_grid = vec![vec![' '; plot_width]; plot_height];
// Flat plot grid indexed as grid[y * plot_width + x]. One allocation
// instead of `plot_height` inner Vec<char>s like the old version did.
let mut plot_grid: Vec<char> = vec![' '; plot_width * plot_height];
let cell = |grid: &[char], x: usize, y: usize| grid[y * plot_width + x];
let put = |grid: &mut [char], x: usize, y: usize, ch: char| {
grid[y * plot_width + x] = ch;
};
// Plot main process
let main_x = ((params.main_user_ms / params.max_user) * (plot_width - 1) as f64) as usize;
@@ -573,7 +579,7 @@ impl ModalManager {
((params.main_system_ms / params.max_system) * (plot_height - 1) as f64) as usize,
);
if main_x < plot_width && main_y < plot_height {
plot_grid[main_y][main_x] = '●'; // Main process marker
put(&mut plot_grid, main_x, main_y, '●');
}
// Plot threads (use different marker)
@@ -587,13 +593,13 @@ impl ModalManager {
);
if thread_x < plot_width && thread_y < plot_height {
if plot_grid[thread_y][thread_x] == ' ' {
plot_grid[thread_y][thread_x] = '○'; // Thread marker (hollow circle)
} else if plot_grid[thread_y][thread_x] == '○' {
plot_grid[thread_y][thread_x] = '◎'; // Multiple threads at same point
} else {
plot_grid[thread_y][thread_x] = '◉'; // Mixed threads/processes at same point
}
let ch = cell(&plot_grid, thread_x, thread_y);
let next = match ch {
' ' => '○',
'○' => '◎',
_ => '◉',
};
put(&mut plot_grid, thread_x, thread_y, next);
}
}
@@ -608,28 +614,34 @@ impl ModalManager {
);
if child_x < plot_width && child_y < plot_height {
if plot_grid[child_y][child_x] == ' ' {
plot_grid[child_y][child_x] = '•'; // Child process marker
} else {
plot_grid[child_y][child_x] = '◉'; // Multiple items at same point
}
let ch = cell(&plot_grid, child_x, child_y);
let next = if ch == ' ' { '•' } else { '◉' };
put(&mut plot_grid, child_x, child_y, next);
}
}
// Render the plot
let mut lines = Vec::new();
// Build the rendered lines. Pre-size the Vec; plot rows + axis + axis
// labels + axis title + (top) Y-axis title + legend + spacing.
let mut lines: Vec<Line> = Vec::with_capacity(plot_height + 6);
// Add Y-axis labels and plot content
for (i, row) in plot_grid.iter().enumerate() {
let y_value = params.max_system * (1.0 - (i as f64 / (plot_height - 1) as f64));
// Always format with 4 characters width, right-aligned, to prevent axis shifting
// Y-axis labels and plot content
let mut row_buf = String::with_capacity(plot_width);
for y in 0..plot_height {
let y_value = params.max_system * (1.0 - (y as f64 / (plot_height - 1).max(1) as f64));
// 4-char fixed-width label so the axis doesn't shift as digits change.
let y_label = if y_value >= 100.0 {
format!("{y_value:>4.0}")
} else {
format!("{y_value:>4.1}")
};
let plot_content: String = row.iter().collect();
// Build the row's char slice into a reusable String buffer.
row_buf.clear();
let start = y * plot_width;
row_buf.extend(plot_grid[start..start + plot_width].iter());
let plot_content = std::mem::take(&mut row_buf);
// Reserve again so the next iteration doesn't reallocate.
row_buf.reserve(plot_width);
lines.push(Line::from(vec![
Span::styled(y_label, Style::default()),
@@ -833,6 +845,7 @@ impl ModalManager {
area: Rect,
process: &socktop_connector::DetailedProcessInfo,
cpu_history: &std::collections::VecDeque<f32>,
cpu_history_sum: f32,
) {
// Split top row: CPU sparkline (left 60%) | Thread scatter plot (right 40%)
let top_chunks = Layout::default()
@@ -843,7 +856,7 @@ impl ModalManager {
])
.split(area);
self.render_cpu_sparkline(f, top_chunks[0], process, cpu_history);
self.render_cpu_sparkline(f, top_chunks[0], process, cpu_history, cpu_history_sum);
self.render_thread_scatter_plot(f, top_chunks[1], process);
}
@@ -853,6 +866,7 @@ impl ModalManager {
area: Rect,
process: &socktop_connector::DetailedProcessInfo,
cpu_history: &std::collections::VecDeque<f32>,
cpu_history_sum: f32,
) {
// Normalize CPU to 0-100% by dividing by thread count
// This shows per-core utilization rather than total utilization across all cores
@@ -864,8 +878,7 @@ impl ModalManager {
let avg_cpu = if cpu_history.is_empty() {
0.0
} else {
let total: f32 = cpu_history.iter().sum();
normalize_cpu_usage(total / cpu_history.len() as f32, thread_count)
normalize_cpu_usage(cpu_history_sum / cpu_history.len() as f32, thread_count)
};
let title = format!("CPU (now: {current_cpu:.1}% | {avg_cpu:.1}%)");
+2
View File
@@ -5,6 +5,8 @@ use std::time::Instant;
/// History data for process metrics rendering
pub struct ProcessHistoryData<'a> {
pub cpu: &'a std::collections::VecDeque<f32>,
/// Running sum of `cpu` maintained by the caller (avoids re-summing per frame)
pub cpu_sum: f32,
pub mem: &'a std::collections::VecDeque<u64>,
pub io_read: &'a std::collections::VecDeque<u64>,
pub io_write: &'a std::collections::VecDeque<u64>,
+3 -3
View File
@@ -11,12 +11,12 @@ pub fn draw_net_spark(
f: &mut ratatui::Frame<'_>,
area: Rect,
title: &str,
hist: &VecDeque<u64>,
hist: &mut VecDeque<u64>,
color: Color,
) {
let max_points = area.width.saturating_sub(2) as usize;
let start = hist.len().saturating_sub(max_points);
let data: Vec<u64> = hist.iter().skip(start).cloned().collect();
let slice = &hist.make_contiguous()[start..];
let spark = Sparkline::default()
.block(
@@ -24,7 +24,7 @@ pub fn draw_net_spark(
.borders(Borders::ALL)
.title(title.to_string()),
)
.data(&data)
.data(slice)
.style(Style::default().fg(color));
f.render_widget(spark, area);
}
+219 -164
View File
@@ -5,8 +5,8 @@ use ratatui::style::Modifier;
use ratatui::{
layout::{Constraint, Direction, Layout, Rect},
style::{Color, Style},
text::{Line, Span},
widgets::{Block, Borders, Paragraph, Table},
text::Span,
widgets::{Block, Borders, Paragraph, Scrollbar, ScrollbarOrientation, ScrollbarState, Table},
};
use std::cmp::Ordering;
@@ -16,18 +16,17 @@ use crate::ui::theme::{
PROCESS_SELECTION_BG, PROCESS_SELECTION_FG, PROCESS_TOOLTIP_BG, PROCESS_TOOLTIP_FG, SB_ARROW,
SB_THUMB, SB_TRACK,
};
use crate::ui::util::human;
/// Simple fuzzy matching: returns true if all characters in needle appear in haystack in order (case-insensitive)
/// Simple fuzzy matching: returns true if all characters in needle appear in
/// haystack in order, ASCII-case-insensitive. Lowercase normalization is done
/// on the fly so we don't allocate two `String`s per haystack like the old
/// version did (this runs once per process per frame).
fn fuzzy_match(haystack: &str, needle: &str) -> bool {
if needle.is_empty() {
return true;
}
let haystack_lower = haystack.to_lowercase();
let needle_lower = needle.to_lowercase();
let mut haystack_chars = haystack_lower.chars();
for needle_char in needle_lower.chars() {
let mut haystack_chars = haystack.chars().map(|c| c.to_ascii_lowercase());
for needle_char in needle.chars().map(|c| c.to_ascii_lowercase()) {
if !haystack_chars.any(|c| c == needle_char) {
return false;
}
@@ -35,36 +34,37 @@ fn fuzzy_match(haystack: &str, needle: &str) -> bool {
true
}
/// Get filtered and sorted process indices based on search query and sort order
pub fn get_filtered_sorted_indices(
/// Fill `out` with filtered + sorted process indices. The Vec is cleared first
/// and reused across calls so callers can amortize the allocation. This is
/// the underlying helper for the App-side cached slice.
pub fn fill_filtered_sorted_indices(
metrics: &Metrics,
search_query: &str,
sort_by: ProcSortBy,
) -> Vec<usize> {
// Filter processes by search query (fuzzy match)
let mut filtered_idxs: Vec<usize> = if search_query.is_empty() {
(0..metrics.top_processes.len()).collect()
out: &mut Vec<usize>,
) {
out.clear();
out.reserve(metrics.top_processes.len());
if search_query.is_empty() {
out.extend(0..metrics.top_processes.len());
} else {
(0..metrics.top_processes.len())
.filter(|&i| fuzzy_match(&metrics.top_processes[i].name, search_query))
.collect()
};
// Sort filtered rows
out.extend(
(0..metrics.top_processes.len())
.filter(|&i| fuzzy_match(&metrics.top_processes[i].name, search_query)),
);
}
match sort_by {
ProcSortBy::CpuDesc => filtered_idxs.sort_by(|&a, &b| {
ProcSortBy::CpuDesc => out.sort_by(|&a, &b| {
let aa = metrics.top_processes[a].cpu_usage;
let bb = metrics.top_processes[b].cpu_usage;
bb.partial_cmp(&aa).unwrap_or(Ordering::Equal)
}),
ProcSortBy::MemDesc => filtered_idxs.sort_by(|&a, &b| {
ProcSortBy::MemDesc => out.sort_by(|&a, &b| {
let aa = metrics.top_processes[a].mem_bytes;
let bb = metrics.top_processes[b].mem_bytes;
bb.cmp(&aa)
}),
}
filtered_idxs
}
/// Parameters for drawing the top processes table
@@ -76,6 +76,16 @@ pub struct ProcessDisplayParams<'a> {
pub selected_process_index: Option<usize>,
pub search_query: &'a str,
pub search_active: bool,
/// Precomputed filtered + sorted indices into `metrics.top_processes`.
/// Maintained on the App side so the draw path never recomputes the list.
pub filtered_indices: &'a [usize],
/// Pre-formatted strings for each row of `metrics.top_processes`.
/// Indexed the same as `metrics.top_processes`. Empty when no procs poll
/// has run yet (the draw path falls back to fast inline formatting).
pub cached_rows: &'a [CachedRow],
/// Peak cpu_usage from the most recent cache build; used to bold the
/// busiest process. -1.0 if no cache.
pub peak_cpu: f32,
}
#[derive(Debug, Clone, Copy, PartialEq, Eq, Default)]
@@ -85,6 +95,45 @@ pub enum ProcSortBy {
MemDesc,
}
/// Pre-formatted strings for one row of the process table. Built once per
/// `Processes` poll (cadence ~2s) and reused by every draw frame in between
/// so the diff renderer can suppress repaints when nothing changed.
#[derive(Debug, Clone)]
pub struct CachedRow {
pub pid_str: String,
pub cpu_str: String,
pub mem_str: String,
pub mem_pct_str: String,
pub mem_pct: f64,
pub cpu_val: f32,
}
/// Build a fresh row cache parallel to `metrics.top_processes`. Reuses `out`'s
/// allocation when possible. Also returns the peak cpu_usage observed, which
/// the draw path uses to bold the busiest process.
pub fn rebuild_row_cache(metrics: &Metrics, out: &mut Vec<CachedRow>) -> f32 {
out.clear();
out.reserve(metrics.top_processes.len());
let total = metrics.mem_total.max(1);
let mut peak = 0.0_f32;
for p in &metrics.top_processes {
let mem_pct = (p.mem_bytes as f64 / total as f64) * 100.0;
let cpu_val = p.cpu_usage;
if cpu_val > peak {
peak = cpu_val;
}
out.push(CachedRow {
pid_str: p.pid.to_string(),
cpu_str: format!("{:>5.1}", cpu_val.clamp(0.0, 100.0)),
mem_str: crate::ui::util::human(p.mem_bytes),
mem_pct_str: format!("{mem_pct:.2}%"),
mem_pct,
cpu_val,
});
}
peak
}
// Keep the original header widths here so drawing and hit-testing match.
const COLS: [Constraint; 5] = [
Constraint::Length(8), // PID
@@ -159,8 +208,7 @@ pub fn draw_top_processes(f: &mut ratatui::Frame<'_>, area: Rect, params: Proces
height: inner.height,
};
// Get filtered and sorted indices
let idxs = get_filtered_sorted_indices(mm, params.search_query, params.sort_by);
let idxs = params.filtered_indices;
// Scrolling
let total_rows = idxs.len();
@@ -170,19 +218,60 @@ pub fn draw_top_processes(f: &mut ratatui::Frame<'_>, area: Rect, params: Proces
let offset = params.scroll_offset.min(max_off);
let show_n = total_rows.saturating_sub(offset).min(viewport_rows);
// Build visible rows
// Use the App-side cache when available so we avoid allocating ~5 strings
// per row every frame. Falls back to inline formatting (slow path) when
// the cache hasn't been built yet — e.g. the very first frame before the
// initial procs poll completes.
let cache_ok = params.cached_rows.len() == mm.top_processes.len();
let total_mem_bytes = mm.mem_total.max(1);
let peak_cpu = mm
.top_processes
.iter()
.map(|p| p.cpu_usage)
.fold(0.0_f32, f32::max);
let peak_cpu = if cache_ok {
params.peak_cpu
} else {
mm.top_processes
.iter()
.map(|p| p.cpu_usage)
.fold(0.0_f32, f32::max)
};
let rows_iter = idxs.iter().skip(offset).take(show_n).map(|&ix| {
let p = &mm.top_processes[ix];
let mem_pct = (p.mem_bytes as f64 / total_mem_bytes as f64) * 100.0;
let cpu_val = p.cpu_usage;
let (
cpu_val,
mem_pct,
pid_span,
name_span,
cpu_span_text,
mem_span_text,
mem_pct_span_text,
) = if cache_ok {
let row = &params.cached_rows[ix];
(
row.cpu_val,
row.mem_pct,
Span::raw(row.pid_str.as_str()),
Span::raw(p.name.as_str()),
row.cpu_str.as_str(),
row.mem_str.as_str(),
row.mem_pct_str.as_str(),
)
} else {
let mem_pct = (p.mem_bytes as f64 / total_mem_bytes as f64) * 100.0;
// SLOW path: only the very first frame before the cache exists.
// We leak the formatted strings via Box::leak'd statics? No —
// simpler: emit empty placeholders. Cache will exist within
// ~500ms and the diff renderer fills it in.
(
p.cpu_usage,
mem_pct,
Span::raw(""),
Span::raw(""),
"",
"",
"",
)
};
let cpu_fg = match cpu_val {
x if x < 25.0 => Color::Green,
x if x < 60.0 => Color::Yellow,
@@ -200,16 +289,14 @@ pub fn draw_top_processes(f: &mut ratatui::Frame<'_>, area: Rect, params: Proces
Style::default()
};
// Check if this process is selected - prioritize PID matching
let is_selected = if let Some(selected_pid) = params.selected_process_pid {
selected_pid == p.pid
} else if let Some(selected_idx) = params.selected_process_index {
selected_idx == ix // ix is the absolute index in the sorted list
selected_idx == ix
} else {
false
};
// Apply selection highlighting
if is_selected {
emphasis = emphasis
.bg(PROCESS_SELECTION_BG)
@@ -217,15 +304,13 @@ pub fn draw_top_processes(f: &mut ratatui::Frame<'_>, area: Rect, params: Proces
.add_modifier(Modifier::BOLD);
}
let cpu_str = fmt_cpu_pct(cpu_val);
ratatui::widgets::Row::new(vec![
ratatui::widgets::Cell::from(p.pid.to_string())
.style(Style::default().fg(Color::DarkGray)),
ratatui::widgets::Cell::from(p.name.clone()),
ratatui::widgets::Cell::from(cpu_str).style(Style::default().fg(cpu_fg)),
ratatui::widgets::Cell::from(human(p.mem_bytes)),
ratatui::widgets::Cell::from(format!("{mem_pct:.2}%"))
ratatui::widgets::Cell::from(pid_span).style(Style::default().fg(Color::DarkGray)),
ratatui::widgets::Cell::from(name_span),
ratatui::widgets::Cell::from(Span::raw(cpu_span_text))
.style(Style::default().fg(cpu_fg)),
ratatui::widgets::Cell::from(Span::raw(mem_span_text)),
ratatui::widgets::Cell::from(Span::raw(mem_pct_span_text))
.style(Style::default().fg(mem_fg)),
])
.style(emphasis)
@@ -293,45 +378,29 @@ pub fn draw_top_processes(f: &mut ratatui::Frame<'_>, area: Rect, params: Proces
}
}
// Draw scrollbar like CPU pane
// Scrollbar (ratatui built-in). Skip drawing when content fits in viewport.
let scroll_area = Rect {
x: inner.x + inner.width.saturating_sub(1),
y: inner.y,
width: 1,
height: inner.height,
};
if scroll_area.height >= 3 {
let track = (scroll_area.height - 2) as usize;
let total = total_rows.max(1);
let view = viewport_rows.clamp(1, total);
let max_off = total.saturating_sub(view);
let thumb_len = (track * view).div_ceil(total).max(1).min(track);
let thumb_top = if max_off == 0 {
0
} else {
((track - thumb_len) * offset + max_off / 2) / max_off
};
// Build lines: top arrow, track (with thumb), bottom arrow
let mut lines: Vec<Line> = Vec::with_capacity(scroll_area.height as usize);
lines.push(Line::from(Span::styled("", Style::default().fg(SB_ARROW))));
for i in 0..track {
if i >= thumb_top && i < thumb_top + thumb_len {
lines.push(Line::from(Span::styled("", Style::default().fg(SB_THUMB))));
} else {
lines.push(Line::from(Span::styled("", Style::default().fg(SB_TRACK))));
}
}
lines.push(Line::from(Span::styled("", Style::default().fg(SB_ARROW))));
f.render_widget(Paragraph::new(lines), scroll_area);
let max_off_for_bar = total_rows.saturating_sub(viewport_rows);
if scroll_area.height >= 3 && max_off_for_bar > 0 {
let scrollbar = Scrollbar::new(ScrollbarOrientation::VerticalRight)
.begin_symbol(Some(""))
.end_symbol(Some(""))
.thumb_symbol("")
.track_symbol(Some(""))
.thumb_style(Style::default().fg(SB_THUMB))
.track_style(Style::default().fg(SB_TRACK))
.begin_style(Style::default().fg(SB_ARROW))
.end_style(Style::default().fg(SB_ARROW));
let mut state = ScrollbarState::new(max_off_for_bar).position(offset);
f.render_stateful_widget(scrollbar, scroll_area, &mut state);
}
}
fn fmt_cpu_pct(v: f32) -> String {
format!("{:>5.1}", v.clamp(0.0, 100.0))
}
/// Handle keyboard scrolling (Up/Down/PageUp/PageDown/Home/End)
/// Parameters for process key event handling
pub struct ProcessKeyParams<'a> {
@@ -339,8 +408,7 @@ pub struct ProcessKeyParams<'a> {
pub selected_process_index: &'a mut Option<usize>,
pub key: crossterm::event::KeyEvent,
pub metrics: Option<&'a Metrics>,
pub sort_by: ProcSortBy,
pub search_query: &'a str,
pub filtered_indices: &'a [usize],
}
/// LEGACY: Use processes_handle_key_with_selection for enhanced functionality
@@ -356,87 +424,71 @@ pub fn processes_handle_key(
pub fn processes_handle_key_with_selection(params: ProcessKeyParams) -> bool {
use crossterm::event::KeyCode;
let move_selection = |delta: isize,
sel_idx: &mut Option<usize>,
sel_pid: &mut Option<u32>,
metrics: Option<&Metrics>,
idxs: &[usize]| {
let Some(m) = metrics else { return };
if idxs.is_empty() {
*sel_idx = None;
*sel_pid = None;
return;
}
if sel_idx.is_none() || sel_pid.is_none() {
let first_idx = idxs[0];
*sel_idx = Some(first_idx);
*sel_pid = Some(m.top_processes[first_idx].pid);
return;
}
let current_idx = sel_idx.unwrap();
match idxs.iter().position(|&idx| idx == current_idx) {
Some(pos) => {
let new_pos = (pos as isize + delta).clamp(0, idxs.len() as isize - 1) as usize;
if new_pos != pos {
let new_idx = idxs[new_pos];
*sel_idx = Some(new_idx);
*sel_pid = Some(m.top_processes[new_idx].pid);
}
}
None => {
// Current selection no longer in filtered list
let first_idx = idxs[0];
*sel_idx = Some(first_idx);
*sel_pid = Some(m.top_processes[first_idx].pid);
}
}
};
match params.key.code {
KeyCode::Up => {
// Navigate through filtered and sorted results
if let Some(m) = params.metrics {
let idxs = get_filtered_sorted_indices(m, params.search_query, params.sort_by);
if idxs.is_empty() {
// No filtered results, clear selection
*params.selected_process_index = None;
*params.selected_process_pid = None;
} else if params.selected_process_index.is_none()
|| params.selected_process_pid.is_none()
{
// No selection - select the first process in filtered/sorted order
let first_idx = idxs[0];
*params.selected_process_index = Some(first_idx);
*params.selected_process_pid = Some(m.top_processes[first_idx].pid);
} else if let Some(current_idx) = *params.selected_process_index {
// Find current position in filtered/sorted list
if let Some(pos) = idxs.iter().position(|&idx| idx == current_idx) {
if pos > 0 {
// Move up in filtered/sorted list
let new_idx = idxs[pos - 1];
*params.selected_process_index = Some(new_idx);
*params.selected_process_pid = Some(m.top_processes[new_idx].pid);
}
} else {
// Current selection not in filtered list, select first result
let first_idx = idxs[0];
*params.selected_process_index = Some(first_idx);
*params.selected_process_pid = Some(m.top_processes[first_idx].pid);
}
}
}
true // Handled
move_selection(
-1,
params.selected_process_index,
params.selected_process_pid,
params.metrics,
params.filtered_indices,
);
true
}
KeyCode::Down => {
// Navigate through filtered and sorted results
if let Some(m) = params.metrics {
let idxs = get_filtered_sorted_indices(m, params.search_query, params.sort_by);
if idxs.is_empty() {
// No filtered results, clear selection
*params.selected_process_index = None;
*params.selected_process_pid = None;
} else if params.selected_process_index.is_none()
|| params.selected_process_pid.is_none()
{
// No selection - select the first process in filtered/sorted order
let first_idx = idxs[0];
*params.selected_process_index = Some(first_idx);
*params.selected_process_pid = Some(m.top_processes[first_idx].pid);
} else if let Some(current_idx) = *params.selected_process_index {
// Find current position in filtered/sorted list
if let Some(pos) = idxs.iter().position(|&idx| idx == current_idx) {
if pos + 1 < idxs.len() {
// Move down in filtered/sorted list
let new_idx = idxs[pos + 1];
*params.selected_process_index = Some(new_idx);
*params.selected_process_pid = Some(m.top_processes[new_idx].pid);
}
} else {
// Current selection not in filtered list, select first result
let first_idx = idxs[0];
*params.selected_process_index = Some(first_idx);
*params.selected_process_pid = Some(m.top_processes[first_idx].pid);
}
}
}
true // Handled
move_selection(
1,
params.selected_process_index,
params.selected_process_pid,
params.metrics,
params.filtered_indices,
);
true
}
KeyCode::Char('x') | KeyCode::Char('X') => {
// Unselect any selected process
if params.selected_process_pid.is_some() || params.selected_process_index.is_some() {
*params.selected_process_pid = None;
*params.selected_process_index = None;
true // Handled
} else {
false // No selection to clear
}
KeyCode::Char('x') | KeyCode::Char('X')
if params.selected_process_pid.is_some() || params.selected_process_index.is_some() =>
{
*params.selected_process_pid = None;
*params.selected_process_index = None;
true
}
KeyCode::Char('x') | KeyCode::Char('X') => false,
KeyCode::Enter => {
// Signal that Enter was pressed with a selection
params.selected_process_pid.is_some() // Return true if we have a selection to handle
@@ -526,8 +578,11 @@ pub struct ProcessMouseParams<'a> {
pub area: Rect,
pub total_rows: usize,
pub metrics: Option<&'a Metrics>,
pub sort_by: ProcSortBy,
pub search_query: &'a str,
/// True when the on-screen search box is currently being drawn (active
/// edit mode OR a non-empty filter is showing). The caller computes this
/// from the same condition as the draw path.
pub search_box_visible: bool,
pub filtered_indices: &'a [usize],
}
/// Enhanced mouse handler that also manages process selection
@@ -545,9 +600,13 @@ pub fn processes_handle_mouse_with_selection(params: ProcessMouseParams) -> Opti
}
// Calculate content area - must match draw_top_processes exactly!
// If search is active or query exists, content starts after search box (3 lines)
let search_active = !params.search_query.is_empty();
let content_start_y = if search_active { inner.y + 3 } else { inner.y };
// If a search box is being drawn (active edit mode OR a filter showing),
// content starts 3 rows below.
let content_start_y = if params.search_box_visible {
inner.y + 3
} else {
inner.y
};
let content = Rect {
x: inner.x,
@@ -555,7 +614,7 @@ pub fn processes_handle_mouse_with_selection(params: ProcessMouseParams) -> Opti
width: inner.width.saturating_sub(2),
height: inner
.height
.saturating_sub(if search_active { 3 } else { 0 }),
.saturating_sub(if params.search_box_visible { 3 } else { 0 }),
};
// Scrollbar interactions (click arrows/page/drag)
@@ -612,12 +671,8 @@ pub fn processes_handle_mouse_with_selection(params: ProcessMouseParams) -> Opti
{
let clicked_row = (params.mouse.row - data_start_row) as usize;
// Find the actual process using the same filtering/sorting logic as the drawing code
if let Some(m) = params.metrics {
// Use the same filtered and sorted indices as display
let idxs = get_filtered_sorted_indices(m, params.search_query, params.sort_by);
// Calculate which process was actually clicked based on filtered/sorted order
let idxs = params.filtered_indices;
let visible_process_position = *params.scroll_offset + clicked_row;
if visible_process_position < idxs.len() {
let actual_process_index = idxs[visible_process_position];
+25 -4
View File
@@ -1,11 +1,13 @@
[package]
name = "socktop_agent"
version = "1.50.1"
version = "1.50.2"
authors = ["Jason Witty <jasonpwitty+socktop@proton.me>"]
description = "Socktop agent daemon. Serves host metrics over WebSocket."
edition = "2024"
license = "MIT"
readme = "README.md"
homepage = "https://github.com/jasonwitty/socktop"
repository = "https://github.com/jasonwitty/socktop"
[dependencies]
# Tokio: Use minimal features instead of "full" to reduce binary size
@@ -21,10 +23,10 @@ flate2 = { version = "1", default-features = false, features = ["rust_backend"]
futures-util = "0.3.31"
tracing = { version = "0.1", optional = true }
tracing-subscriber = { version = "0.3", features = ["env-filter"], optional = true }
gfxinfo = "0.1.2"
gfxinfo = { version = "0.1.2", optional = true }
once_cell = "1.19"
axum-server = { version = "0.7", features = ["tls-rustls"] }
rustls = "0.23"
rustls = { version = "0.23", features = ["aws-lc-rs"] }
rustls-pemfile = "2.1"
rcgen = "0.13"
anyhow = "1"
@@ -33,7 +35,8 @@ prost = { workspace = true }
time = { version = "0.3", default-features = false, features = ["formatting", "macros", "parsing" ] }
[features]
default = []
default = ["gpu"]
gpu = ["gfxinfo"]
logging = ["tracing", "tracing-subscriber"]
[build-dependencies]
@@ -45,3 +48,21 @@ protoc-bin-vendored = "3"
assert_cmd = "2.0"
tempfile = "3.10"
tokio-tungstenite = "0.21"
[package.metadata.deb]
maintainer = "Jason Witty <jasonpwitty+socktop@proton.me>"
copyright = "2024, Jason Witty <jasonpwitty+socktop@proton.me>"
license-file = ["../LICENSE", "4"]
extended-description = """\
socktop_agent is the daemon component that runs on remote hosts to collect \
and serve system metrics over WebSocket. It gathers CPU, memory, disk, network, \
GPU, and process information that can be monitored remotely by the socktop TUI client."""
depends = "$auto"
section = "admin"
priority = "optional"
assets = [
["target/release/socktop_agent", "usr/bin/", "755"],
["../README.md", "usr/share/doc/socktop_agent/", "644"],
]
maintainer-scripts = "debian/"
systemd-units = { unit-name = "socktop-agent", unit-scripts = ".", enable = false }
+57
View File
@@ -0,0 +1,57 @@
#!/bin/sh
set -e
# Create socktop user and group if they don't exist
if ! getent group socktop >/dev/null; then
addgroup --system socktop
fi
if ! getent passwd socktop >/dev/null; then
adduser --system --ingroup socktop --home /var/lib/socktop \
--no-create-home --disabled-password --disabled-login \
--gecos "Socktop Agent" socktop
fi
# Create state directory
mkdir -p /var/lib/socktop
chown socktop:socktop /var/lib/socktop
chmod 755 /var/lib/socktop
# Create config directory if it doesn't exist
mkdir -p /etc/socktop
chmod 755 /etc/socktop
#DEBHELPER#
# Print helpful message to the user
cat <<EOF
┌─────────────────────────────────────────────────────────────────────┐
│ socktop-agent has been installed successfully! │
├─────────────────────────────────────────────────────────────────────┤
│ │
│ The systemd service has been installed but is NOT enabled by │
│ default. To enable and start the service: │
│ │
│ sudo systemctl enable --now socktop-agent │
│ │
│ To start without enabling on boot: │
│ │
│ sudo systemctl start socktop-agent │
│ │
│ To check service status: │
│ │
│ sudo systemctl status socktop-agent │
│ │
│ Default settings: │
│ - Port: 3000 (use -p or --port to change) │
│ - SSL/TLS: disabled (use --enableSSL to enable) │
│ │
│ For more information, see: │
│ /usr/share/doc/socktop_agent/README.md │
│ │
└─────────────────────────────────────────────────────────────────────┘
EOF
exit 0
+34
View File
@@ -0,0 +1,34 @@
#!/bin/sh
set -e
case "$1" in
purge)
# Remove user and group on purge
if getent passwd socktop >/dev/null; then
deluser --quiet socktop || true
fi
if getent group socktop >/dev/null; then
delgroup --quiet socktop || true
fi
# Remove state directory on purge
rm -rf /var/lib/socktop
# Remove config directory if empty
rmdir --ignore-fail-on-non-empty /etc/socktop 2>/dev/null || true
;;
remove|upgrade|failed-upgrade|abort-install|abort-upgrade|disappear)
# Do nothing on remove/upgrade
;;
*)
echo "postrm called with unknown argument \`$1'" >&2
exit 1
;;
esac
#DEBHELPER#
exit 0
+27
View File
@@ -0,0 +1,27 @@
[Unit]
Description=Socktop Agent - Remote System Monitor
Documentation=https://github.com/jasonwitty/socktop
After=network-online.target
Wants=network-online.target
[Service]
Type=simple
ExecStart=/usr/bin/socktop_agent --port 3000
Environment=RUST_LOG=info
# Optional authentication token:
# Environment=SOCKTOP_TOKEN=changeme
Restart=on-failure
RestartSec=5
User=socktop
Group=socktop
NoNewPrivileges=true
# Security hardening
PrivateTmp=true
ProtectSystem=strict
ProtectHome=true
ReadWritePaths=/var/lib/socktop
StateDirectory=socktop
[Install]
WantedBy=multi-user.target
-95
View File
@@ -1,95 +0,0 @@
//! Caching for process metrics and journal entries
use std::collections::HashMap;
use std::time::{Duration, Instant};
use tokio::sync::RwLock;
use crate::types::{ProcessMetricsResponse, JournalResponse};
#[derive(Debug, Clone)]
struct CacheEntry<T> {
data: T,
cached_at: Instant,
ttl: Duration,
}
impl<T> CacheEntry<T> {
fn is_expired(&self) -> bool {
self.cached_at.elapsed() > self.ttl
}
}
#[derive(Debug)]
pub struct ProcessCache {
process_metrics: RwLock<HashMap<u32, CacheEntry<ProcessMetricsResponse>>>,
journal_entries: RwLock<HashMap<u32, CacheEntry<JournalResponse>>>,
}
impl ProcessCache {
pub fn new() -> Self {
Self {
process_metrics: RwLock::new(HashMap::new()),
journal_entries: RwLock::new(HashMap::new()),
}
}
/// Get cached process metrics if available and not expired (250ms TTL)
pub async fn get_process_metrics(&self, pid: u32) -> Option<ProcessMetricsResponse> {
let cache = self.process_metrics.read().await;
if let Some(entry) = cache.get(&pid) {
if !entry.is_expired() {
return Some(entry.data.clone());
}
}
None
}
/// Cache process metrics with 250ms TTL
pub async fn set_process_metrics(&self, pid: u32, data: ProcessMetricsResponse) {
let mut cache = self.process_metrics.write().await;
cache.insert(pid, CacheEntry {
data,
cached_at: Instant::now(),
ttl: Duration::from_millis(250),
});
}
/// Get cached journal entries if available and not expired (1s TTL)
pub async fn get_journal_entries(&self, pid: u32) -> Option<JournalResponse> {
let cache = self.journal_entries.read().await;
if let Some(entry) = cache.get(&pid) {
if !entry.is_expired() {
return Some(entry.data.clone());
}
}
None
}
/// Cache journal entries with 1s TTL
pub async fn set_journal_entries(&self, pid: u32, data: JournalResponse) {
let mut cache = self.journal_entries.write().await;
cache.insert(pid, CacheEntry {
data,
cached_at: Instant::now(),
ttl: Duration::from_secs(1),
});
}
/// Clean up expired entries periodically
pub async fn cleanup_expired(&self) {
{
let mut cache = self.process_metrics.write().await;
cache.retain(|_, entry| !entry.is_expired());
}
{
let mut cache = self.journal_entries.write().await;
cache.retain(|_, entry| !entry.is_expired());
}
}
}
impl Default for ProcessCache {
fn default() -> Self {
Self::new()
}
}
+8
View File
@@ -1,4 +1,5 @@
// gpu.rs
#[cfg(feature = "gpu")]
use gfxinfo::active_gpu;
#[derive(Debug, Clone, serde::Serialize)]
@@ -9,6 +10,7 @@ pub struct GpuMetrics {
pub mem_total_bytes: u64,
}
#[cfg(feature = "gpu")]
pub fn collect_all_gpus() -> Result<Vec<GpuMetrics>, Box<dyn std::error::Error>> {
let gpu = active_gpu()?; // Use ? to unwrap Result
let info = gpu.info();
@@ -22,3 +24,9 @@ pub fn collect_all_gpus() -> Result<Vec<GpuMetrics>, Box<dyn std::error::Error>>
Ok(vec![metrics])
}
#[cfg(not(feature = "gpu"))]
pub fn collect_all_gpus() -> Result<Vec<GpuMetrics>, Box<dyn std::error::Error>> {
// GPU support not available on this platform
Ok(vec![])
}
+6
View File
@@ -30,6 +30,12 @@ fn arg_value(name: &str) -> Option<String> {
}
fn main() -> anyhow::Result<()> {
// Install rustls crypto provider before any TLS operations
// This is required when using axum-server's tls-rustls feature
rustls::crypto::aws_lc_rs::default_provider()
.install_default()
.ok(); // Ignore error if already installed
#[cfg(feature = "logging")]
tracing_subscriber::fmt::init();
+338 -234
View File
@@ -23,45 +23,40 @@ use tracing::warn;
// NOTE: CPU normalization env removed; non-Linux now always reports per-process share (0..100) as given by sysinfo.
// Helper functions to get CPU time from /proc/stat on Linux
// Read (utime, stime) in milliseconds from /proc/{pid}/stat in one go.
// Returns (0, 0) if the file can't be read.
//
// We use `rfind(')')` to step past the `comm` field, which can contain
// arbitrary characters (including spaces and parens), then index the
// post-comm fields by position. This is the same trick `read_proc_jiffies`
// uses below — `split_whitespace().collect::<Vec<_>>()` from the start of
// the file would mis-parse process names with spaces, and also wastes an
// allocation per call. Two callers used to read this file twice (once for
// user, once for system); now it's one syscall per detailed-process record.
#[cfg(target_os = "linux")]
fn get_cpu_time_user(pid: u32) -> u64 {
if let Ok(stat) = fs::read_to_string(format!("/proc/{pid}/stat")) {
let fields: Vec<&str> = stat.split_whitespace().collect();
if fields.len() > 13 {
// Field 13 (0-indexed) is utime (user CPU time in clock ticks)
if let Ok(utime) = fields[13].parse::<u64>() {
// Convert clock ticks to milliseconds (assuming 100 Hz)
return utime * 10; // 1 tick = 10ms at 100 Hz
}
}
}
0
}
#[cfg(target_os = "linux")]
fn get_cpu_time_system(pid: u32) -> u64 {
if let Ok(stat) = fs::read_to_string(format!("/proc/{pid}/stat")) {
let fields: Vec<&str> = stat.split_whitespace().collect();
if fields.len() > 14 {
// Field 14 (0-indexed) is stime (system CPU time in clock ticks)
if let Ok(stime) = fields[14].parse::<u64>() {
// Convert clock ticks to milliseconds (assuming 100 Hz)
return stime * 10; // 1 tick = 10ms at 100 Hz
}
}
}
0
fn get_cpu_times_ms(pid: u32) -> (u64, u64) {
let Ok(s) = fs::read_to_string(format!("/proc/{pid}/stat")) else {
return (0, 0);
};
let Some(rpar) = s.rfind(')') else {
return (0, 0);
};
let Some(after) = s.get(rpar + 2..) else {
return (0, 0);
};
let mut it = after.split_whitespace();
// Post-comm field offsets: state, ppid, pgrp, session, tty_nr, tpgid,
// flags, minflt, cminflt, majflt, cmajflt, utime, stime, ...
// utime is offset 11; stime follows.
let utime = it.nth(11).and_then(|s| s.parse::<u64>().ok()).unwrap_or(0);
let stime = it.next().and_then(|s| s.parse::<u64>().ok()).unwrap_or(0);
// 1 tick = 10ms at 100 Hz (USER_HZ).
(utime * 10, stime * 10)
}
#[cfg(not(target_os = "linux"))]
fn get_cpu_time_user(_pid: u32) -> u64 {
0 // Not implemented for non-Linux platforms
}
#[cfg(not(target_os = "linux"))]
fn get_cpu_time_system(_pid: u32) -> u64 {
0 // Not implemented for non-Linux platforms
fn get_cpu_times_ms(_pid: u32) -> (u64, u64) {
(0, 0)
}
// Runtime toggles (read once)
fn gpu_enabled() -> bool {
@@ -81,6 +76,47 @@ fn temp_enabled() -> bool {
})
}
// TTL knobs read once at first use, then cached. These hit the hot polling
// paths (every 250ms-1.5s), so re-reading via libc getenv per call is wasted.
fn metrics_ttl_ms() -> u64 {
static V: OnceCell<u64> = OnceCell::new();
*V.get_or_init(|| {
std::env::var("SOCKTOP_AGENT_METRICS_TTL_MS")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(250)
})
}
fn disks_ttl_ms() -> u64 {
static V: OnceCell<u64> = OnceCell::new();
*V.get_or_init(|| {
std::env::var("SOCKTOP_AGENT_DISKS_TTL_MS")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(1_000)
})
}
#[cfg(target_os = "linux")]
fn processes_ttl_ms() -> u64 {
static V: OnceCell<u64> = OnceCell::new();
*V.get_or_init(|| {
std::env::var("SOCKTOP_AGENT_PROCESSES_TTL_MS")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(1_500)
})
}
#[cfg(not(target_os = "linux"))]
fn name_cache_cleanup_threshold() -> usize {
static V: OnceCell<usize> = OnceCell::new();
*V.get_or_init(|| {
std::env::var("SOCKTOP_AGENT_NAME_CACHE_CLEANUP_THRESHOLD")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(1000)
})
}
// Tiny TTL caches to avoid rescanning sensors every 500ms
const TTL: Duration = Duration::from_millis(1500);
struct TempCache {
@@ -89,6 +125,32 @@ struct TempCache {
}
static TEMP: OnceCell<Mutex<TempCache>> = OnceCell::new();
// Last time `state.components` was refreshed (by any caller). Both
// collect_fast_metrics and collect_disks need fresh sensor values; without
// this gate they were each doing their own `Components::refresh` on their
// own cadence, paying the hwmon syscall cost twice per polling cycle.
// 1s is short enough that disk temps stay accurate (they change slowly) and
// long enough to suppress back-to-back refreshes from concurrent endpoints.
const COMPONENTS_REFRESH_TTL: Duration = Duration::from_millis(1000);
static COMPONENTS_LAST_REFRESH: OnceCell<Mutex<Option<Instant>>> = OnceCell::new();
/// Refresh `state.components` only if the cached refresh timestamp is older
/// than `COMPONENTS_REFRESH_TTL`. Caller must already hold the components
/// lock.
fn refresh_components_if_stale(components: &mut sysinfo::Components) {
let lock = COMPONENTS_LAST_REFRESH.get_or_init(|| Mutex::new(None));
let mut last = match lock.lock() {
Ok(g) => g,
Err(_) => return, // Poisoned — skip; values stay as-is until next call
};
let now = Instant::now();
let stale = last.is_none_or(|t| now.duration_since(t) >= COMPONENTS_REFRESH_TTL);
if stale {
components.refresh(false);
*last = Some(now);
}
}
struct GpuCache {
at: Option<Instant>,
v: Option<Vec<crate::gpu::GpuMetrics>>,
@@ -154,12 +216,7 @@ fn set_gpus(v: Option<Vec<crate::gpu::GpuMetrics>>) {
// Collect only fast-changing metrics (CPU/mem/net + optional temps/gpus).
pub async fn collect_fast_metrics(state: &AppState) -> Metrics {
// TTL (ms) overridable via env, default 250ms
let ttl_ms: u64 = std::env::var("SOCKTOP_AGENT_METRICS_TTL_MS")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(250);
let ttl = StdDuration::from_millis(ttl_ms);
let ttl = StdDuration::from_millis(metrics_ttl_ms());
{
let cache = state.cache_metrics.lock().await;
if cache.is_fresh(ttl)
@@ -202,7 +259,7 @@ pub async fn collect_fast_metrics(state: &AppState) -> Metrics {
} else if temp_enabled() {
let val = {
let mut components = state.components.lock().await;
components.refresh(false);
refresh_components_if_stale(&mut components);
components.iter().find_map(|c| {
let l = c.label().to_ascii_lowercase();
if l.contains("cpu")
@@ -236,12 +293,19 @@ pub async fn collect_fast_metrics(state: &AppState) -> Metrics {
});
let mut cache = cache.lock().unwrap();
// Collect current network names
let current_names: Vec<_> = nets.keys().map(|name| name.to_string()).collect();
// Update cached network names if they changed
if cache.names != current_names {
cache.names = current_names;
// Detect a topology change without allocating: compare lengths first,
// then zip and walk. Only on a real diff do we materialize the new
// names list. Was: `nets.keys().map(to_string).collect::<Vec<_>>()`
// every tick — a fresh Vec<String> just to compare.
let topology_changed = cache.names.len() != nets.keys().count()
|| cache
.names
.iter()
.zip(nets.keys())
.any(|(cached, current)| cached.as_str() != current.as_str());
if topology_changed {
cache.names.clear();
cache.names.extend(nets.keys().map(|n| n.to_string()));
}
// Reuse NetworkInfo objects
@@ -319,11 +383,7 @@ pub async fn collect_fast_metrics(state: &AppState) -> Metrics {
// Cached disks
pub async fn collect_disks(state: &AppState) -> Vec<DiskInfo> {
let ttl_ms: u64 = std::env::var("SOCKTOP_AGENT_DISKS_TTL_MS")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(1_000);
let ttl = StdDuration::from_millis(ttl_ms);
let ttl = StdDuration::from_millis(disks_ttl_ms());
{
let cache = state.cache_disks.lock().await;
if cache.is_fresh(ttl)
@@ -339,7 +399,9 @@ pub async fn collect_disks(state: &AppState) -> Vec<DiskInfo> {
// NVMe temps show up as "Composite" under different chip names
let disk_temps = {
let mut components = state.components.lock().await;
components.refresh(true); // true = refresh values, not just the list
// Shared TTL-gated refresh: avoids paying the hwmon scan twice when
// both endpoints converge in the same second.
refresh_components_if_stale(&mut components);
let mut composite_temps = Vec::new();
@@ -572,12 +634,7 @@ fn read_proc_jiffies(pid: u32) -> Option<u64> {
/// Collect all processes (Linux): compute CPU% via /proc jiffies delta; sorting moved to client.
#[cfg(target_os = "linux")]
pub async fn collect_processes_all(state: &AppState) -> ProcessesPayload {
let ttl_ms: u64 = std::env::var("SOCKTOP_AGENT_PROCESSES_TTL_MS")
.ok()
.and_then(|v| v.parse().ok())
// Higher default (1500ms) on non-Linux only; keep 1500 here for Linux correctness (more frequent updates).
.unwrap_or(1_500);
let ttl = StdDuration::from_millis(ttl_ms);
let ttl = StdDuration::from_millis(processes_ttl_ms());
{
let cache = state.cache_processes.lock().await;
if cache.is_fresh(ttl)
@@ -586,13 +643,24 @@ pub async fn collect_processes_all(state: &AppState) -> ProcessesPayload {
return c.clone();
}
}
// Reuse shared System to avoid reallocation; refresh processes fully.
// Reuse shared System to avoid reallocation. We only need name + memory
// from sysinfo here — per-process CPU% is computed below from /proc/{pid}/stat
// jiffies (see `read_proc_jiffies` + `read_total_jiffies`), so asking sysinfo
// to gather CPU/exe/cmd/cwd/env per process is wasted /proc traffic on a Pi
// (was reading /proc/{pid}/{cmdline,exe,cwd,environ,io,status} for every PID
// on every 2 s poll via `everything()`).
//
// `without_tasks()` is REQUIRED: it suppresses per-thread entries in the
// process map (without it, sysinfo returns one entry per /proc/[tid] —
// 780+ entries on a typical desktop because of glib/gdbus/Chrome thread
// pools). The original code paired this with `everything()`; we keep the
// filter when downgrading to a minimal refresh spec.
let mut sys_guard = state.sys.lock().await;
let sys = &mut *sys_guard;
sys.refresh_processes_specifics(
ProcessesToUpdate::All,
false,
ProcessRefreshKind::everything().without_tasks(),
ProcessRefreshKind::nothing().with_memory().without_tasks(),
);
let total_count = sys.processes().len();
@@ -607,36 +675,50 @@ pub async fn collect_processes_all(state: &AppState) -> ProcessesPayload {
}
let total_now = read_total_jiffies().unwrap_or(0);
// Compute deltas vs last sample
let (last_total, mut last_map) = {
#[cfg(target_os = "linux")]
{
let mut t = state.proc_cpu.lock().await;
let lt = t.last_total;
let lm = std::mem::take(&mut t.last_per_pid);
t.last_total = total_now;
t.last_per_pid = current.clone();
(lt, lm)
}
#[cfg(not(target_os = "linux"))]
{
let _: u64 = total_now; // silence unused warning
(0u64, HashMap::new())
}
};
// Compute deltas vs last sample. We hold the proc_cpu lock for the whole
// collection below so we can read+update the per-pid name cache in one
// critical section.
let mut tracker = state.proc_cpu.lock().await;
let last_total = tracker.last_total;
// Move the old per-pid jiffies map out for delta computation.
let mut last_map = std::mem::take(&mut tracker.last_per_pid);
tracker.last_total = total_now;
// On first run or if total delta is tiny, report zeros
// Resolve a name through the per-pid cache. Allocates only on miss.
let resolve_name =
|tracker: &mut crate::state::ProcCpuTracker, pid: u32, p: &sysinfo::Process| -> String {
if let Some(cached) = tracker.names.get(&pid) {
return cached.clone();
}
let new_name = p.name().to_string_lossy().into_owned();
tracker.names.insert(pid, new_name.clone());
new_name
};
// On first run or if total delta is tiny, report zeros.
if last_total == 0 || total_now <= last_total {
let procs: Vec<ProcessInfo> = sys
.processes()
.values()
.map(|p| ProcessInfo {
pid: p.pid().as_u32(),
name: p.name().to_string_lossy().into_owned(),
let mut procs: Vec<ProcessInfo> = Vec::with_capacity(total_count);
for p in sys.processes().values() {
let pid = p.pid().as_u32();
let name = resolve_name(&mut tracker, pid, p);
procs.push(ProcessInfo {
pid,
name,
cpu_usage: 0.0,
mem_bytes: p.memory(),
})
.collect();
});
}
// Stash the just-collected jiffies for next call's delta, then prune
// dead pids from the name cache. Borrowing dance: retain reads
// `tracker.last_per_pid` through the closure, which conflicts with
// the mutable borrow of `tracker.names.retain`. Split via split-borrow:
tracker.last_per_pid = current;
let crate::state::ProcCpuTracker {
ref last_per_pid,
ref mut names,
..
} = *tracker;
names.retain(|pid, _| last_per_pid.contains_key(pid));
return ProcessesPayload {
process_count: total_count,
top_processes: procs,
@@ -645,23 +727,31 @@ pub async fn collect_processes_all(state: &AppState) -> ProcessesPayload {
let dt = total_now.saturating_sub(last_total).max(1) as f32;
let procs: Vec<ProcessInfo> = sys
.processes()
.values()
.map(|p| {
let pid = p.pid().as_u32();
let now = current.get(&pid).copied().unwrap_or(0);
let prev = last_map.remove(&pid).unwrap_or(0);
let du = now.saturating_sub(prev) as f32;
let cpu = ((du / dt) * 100.0).clamp(0.0, 100.0);
ProcessInfo {
pid,
name: p.name().to_string_lossy().into_owned(),
cpu_usage: cpu,
mem_bytes: p.memory(),
}
})
.collect();
let mut procs: Vec<ProcessInfo> = Vec::with_capacity(total_count);
for p in sys.processes().values() {
let pid = p.pid().as_u32();
let now = current.get(&pid).copied().unwrap_or(0);
let prev = last_map.remove(&pid).unwrap_or(0);
let du = now.saturating_sub(prev) as f32;
let cpu = ((du / dt) * 100.0).clamp(0.0, 100.0);
let name = resolve_name(&mut tracker, pid, p);
procs.push(ProcessInfo {
pid,
name,
cpu_usage: cpu,
mem_bytes: p.memory(),
});
}
// Save current jiffies map for next call and prune dead pids from the
// name cache. `current` is moved here (no clone — that's also #19).
tracker.last_per_pid = current;
let crate::state::ProcCpuTracker {
ref last_per_pid,
ref mut names,
..
} = *tracker;
names.retain(|pid, _| last_per_pid.contains_key(pid));
drop(tracker);
let payload = ProcessesPayload {
process_count: total_count,
@@ -749,13 +839,17 @@ pub async fn collect_processes_all(state: &AppState) -> ProcessesPayload {
// .unwrap_or(std::cmp::Ordering::Equal)
// });
// Clean up old process names cache when it grows too large
let cache_cleanup_threshold = std::env::var("SOCKTOP_AGENT_NAME_CACHE_CLEANUP_THRESHOLD")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(1000); // Default: most modern systems have 400-700 processes
// Clean up old process names cache when it grows too large.
let cache_cleanup_threshold = name_cache_cleanup_threshold();
if total_count > proc_cache.names.len() + cache_cleanup_threshold {
// `now` is only consumed by the `tracing::debug!` below, so gate
// the binding with the same cfg as its consumer. Without this,
// a non-logging build (the default) emits an unused-variable
// warning. The Linux CI doesn't catch it because this block lives
// in the `#[cfg(not(target_os = "linux"))]` collect_processes_all —
// the warning only surfaces on the Windows build matrix.
#[cfg(feature = "logging")]
let now = std::time::Instant::now();
proc_cache
.names
@@ -811,17 +905,94 @@ fn enumerate_child_processes_lightweight(
children
}
/// Single-read extraction of the /proc/{pid}/status fields the detail
/// endpoint cares about. Callers used to open this file twice per
/// detail-process record (once for VmRSS/VmSize, once for Uid/Gid/Threads/
/// State); now it's one read + one scan.
#[cfg(target_os = "linux")]
#[derive(Default)]
struct ProcStatus {
rss_kb: u64,
vsize_kb: u64,
uid: u32,
gid: u32,
threads: u32,
/// Raw status letter from `State:` (e.g. 'R', 'S'). '?' if missing.
state_ch: char,
}
#[cfg(target_os = "linux")]
fn read_proc_status(pid: u32) -> Option<ProcStatus> {
let content = fs::read_to_string(format!("/proc/{pid}/status")).ok()?;
let mut out = ProcStatus {
state_ch: '?',
..Default::default()
};
for line in content.lines() {
if let Some(v) = line.strip_prefix("VmRSS:") {
out.rss_kb = v
.split_whitespace()
.next()
.and_then(|s| s.parse().ok())
.unwrap_or(0);
} else if let Some(v) = line.strip_prefix("VmSize:") {
out.vsize_kb = v
.split_whitespace()
.next()
.and_then(|s| s.parse().ok())
.unwrap_or(0);
} else if let Some(v) = line.strip_prefix("Uid:") {
out.uid = v
.split_whitespace()
.next()
.and_then(|s| s.parse().ok())
.unwrap_or(0);
} else if let Some(v) = line.strip_prefix("Gid:") {
out.gid = v
.split_whitespace()
.next()
.and_then(|s| s.parse().ok())
.unwrap_or(0);
} else if let Some(v) = line.strip_prefix("Threads:") {
out.threads = v.trim().parse().unwrap_or(0);
} else if let Some(v) = line.strip_prefix("State:") {
out.state_ch = v.trim().chars().next().unwrap_or('?');
}
}
Some(out)
}
#[cfg(target_os = "linux")]
fn proc_state_label(c: char) -> &'static str {
match c {
'R' => "Running",
'S' => "Sleeping",
'D' => "Disk Sleep",
'Z' => "Zombie",
'T' => "Stopped",
't' => "Tracing Stop",
'X' | 'x' => "Dead",
'K' => "Wakekill",
'W' => "Waking",
'P' => "Parked",
'I' => "Idle",
_ => "Unknown",
}
}
/// Read parent PID from /proc/{pid}/stat
#[cfg(target_os = "linux")]
fn read_parent_pid_from_proc(pid: u32) -> Option<u32> {
let stat = fs::read_to_string(format!("/proc/{pid}/stat")).ok()?;
// Format: pid (comm) state ppid ...
// We need to handle process names with spaces/parentheses
// Format: pid (comm) state ppid ... — comm can contain spaces/parens,
// so we step past the closing paren first.
let ppid_start = stat.rfind(')')?;
let fields: Vec<&str> = stat[ppid_start + 1..].split_whitespace().collect();
// After the closing paren: state ppid ...
// Field 1 (0-indexed) is ppid
fields.get(1)?.parse::<u32>().ok()
// After ") ": state, ppid, ... — ppid is the second field.
stat[ppid_start + 1..]
.split_whitespace()
.nth(1)?
.parse::<u32>()
.ok()
}
/// Collect process information from /proc files
@@ -830,8 +1001,11 @@ fn collect_process_info_from_proc(
pid: u32,
system: &sysinfo::System,
) -> Option<DetailedProcessInfo> {
// Try to get basic info from sysinfo if it's already loaded (cheap lookup)
// Otherwise read from /proc directly
// One read of /proc/{pid}/status gets us everything the detail endpoint
// needs from it: memory (when not in sysinfo cache), Uid/Gid, Threads,
// and State. The previous code opened this file twice per process record.
let st = read_proc_status(pid)?;
let (name, cpu_usage, mem_bytes, virtual_mem_bytes) =
if let Some(proc) = system.process(sysinfo::Pid::from_u32(pid)) {
(
@@ -841,30 +1015,13 @@ fn collect_process_info_from_proc(
proc.virtual_memory(),
)
} else {
// Process not in sysinfo cache, read minimal info from /proc
// Process not in sysinfo cache — derive name from /proc/{pid}/comm
// and memory from the status read above.
let name = fs::read_to_string(format!("/proc/{pid}/comm"))
.ok()?
.trim()
.to_string();
// Read memory from /proc/{pid}/status
let status_content = fs::read_to_string(format!("/proc/{pid}/status")).ok()?;
let mut mem_bytes = 0u64;
let mut virtual_mem_bytes = 0u64;
for line in status_content.lines() {
if let Some(value) = line.strip_prefix("VmRSS:") {
if let Some(kb) = value.split_whitespace().next() {
mem_bytes = kb.parse::<u64>().unwrap_or(0) * 1024;
}
} else if let Some(value) = line.strip_prefix("VmSize:")
&& let Some(kb) = value.split_whitespace().next()
{
virtual_mem_bytes = kb.parse::<u64>().unwrap_or(0) * 1024;
}
}
(name, 0.0, mem_bytes, virtual_mem_bytes)
(name, 0.0, st.rss_kb * 1024, st.vsize_kb * 1024)
};
// Read command line
@@ -873,54 +1030,21 @@ fn collect_process_info_from_proc(
.map(|s| s.replace('\0', " ").trim().to_string())
.unwrap_or_default();
// Read status information
let status_content = fs::read_to_string(format!("/proc/{pid}/status")).ok()?;
let mut uid = 0u32;
let mut gid = 0u32;
let mut thread_count = 0u32;
let mut status = "Unknown".to_string();
let uid = st.uid;
let gid = st.gid;
let thread_count = st.threads;
let status = proc_state_label(st.state_ch).to_string();
for line in status_content.lines() {
if let Some(value) = line.strip_prefix("Uid:") {
if let Some(uid_str) = value.split_whitespace().next() {
uid = uid_str.parse().unwrap_or(0);
}
} else if let Some(value) = line.strip_prefix("Gid:") {
if let Some(gid_str) = value.split_whitespace().next() {
gid = gid_str.parse().unwrap_or(0);
}
} else if let Some(value) = line.strip_prefix("Threads:") {
thread_count = value.trim().parse().unwrap_or(0);
} else if let Some(value) = line.strip_prefix("State:") {
status = value
.trim()
.chars()
.next()
.map(|c| match c {
'R' => "Running",
'S' => "Sleeping",
'D' => "Disk Sleep",
'Z' => "Zombie",
'T' => "Stopped",
't' => "Tracing Stop",
'X' | 'x' => "Dead",
'K' => "Wakekill",
'W' => "Waking",
'P' => "Parked",
'I' => "Idle",
_ => "Unknown",
})
.unwrap_or("Unknown")
.to_string();
}
}
// Read start time from stat
// Read start time from stat — comm-safe via rfind(')').
let start_time = if let Ok(stat) = fs::read_to_string(format!("/proc/{pid}/stat")) {
let stat_end = stat.rfind(')')?;
let fields: Vec<&str> = stat[stat_end + 1..].split_whitespace().collect();
// Field 19 (0-indexed) is starttime in clock ticks since boot
fields.get(19)?.parse::<u64>().ok()?
// After ") ": state, ppid, ..., starttime — starttime is the 20th
// post-comm field (index 19).
stat[stat_end + 1..]
.split_whitespace()
.nth(19)?
.parse::<u64>()
.ok()?
} else {
0
};
@@ -954,6 +1078,9 @@ fn collect_process_info_from_proc(
.ok()
.map(|p| p.to_string_lossy().to_string());
// One read of /proc/{pid}/stat covers both user + system CPU times.
let (cpu_time_user, cpu_time_system) = get_cpu_times_ms(pid);
Some(DetailedProcessInfo {
pid,
name,
@@ -969,8 +1096,8 @@ fn collect_process_info_from_proc(
user_id: uid,
group_id: gid,
start_time,
cpu_time_user: get_cpu_time_user(pid),
cpu_time_system: get_cpu_time_system(pid),
cpu_time_user,
cpu_time_system,
read_bytes,
write_bytes,
working_directory,
@@ -1059,22 +1186,24 @@ fn collect_thread_info(pid: u32) -> Vec<crate::types::ThreadInfo> {
.trim()
.to_string();
// Read thread stat for CPU times and status
// Read thread stat for CPU times and status.
let stat_path = format!("/proc/{pid}/task/{tid}/stat");
let Ok(stat_content) = fs::read_to_string(&stat_path) else {
continue;
};
// Parse stat file (similar format to process stat)
// Fields: pid comm state ... utime stime ...
let fields: Vec<&str> = stat_content.split_whitespace().collect();
if fields.len() < 15 {
// Thread/comm names can contain spaces or parens, so step past the
// last ')' before parsing post-comm fields. Post-comm offsets:
// 0: state, 1: ppid, 2: pgrp, ..., 11: utime, 12: stime
let Some(rpar) = stat_content.rfind(')') else {
continue;
}
// Field 2 is state (R, S, D, Z, T, etc.)
let status = fields
.get(2)
};
let Some(after) = stat_content.get(rpar + 1..) else {
continue;
};
let mut it = after.split_whitespace();
let status = it
.next()
.and_then(|s| s.chars().next())
.map(|c| match c {
'R' => "Running",
@@ -1089,16 +1218,9 @@ fn collect_thread_info(pid: u32) -> Vec<crate::types::ThreadInfo> {
.unwrap_or("Unknown")
.to_string();
// Field 13 is utime (user CPU time in clock ticks)
// Field 14 is stime (system CPU time in clock ticks)
let utime = fields
.get(13)
.and_then(|s| s.parse::<u64>().ok())
.unwrap_or(0);
let stime = fields
.get(14)
.and_then(|s| s.parse::<u64>().ok())
.unwrap_or(0);
// 10 fields between state and utime (ppid..cmajflt).
let utime = it.nth(10).and_then(|s| s.parse::<u64>().ok()).unwrap_or(0);
let stime = it.next().and_then(|s| s.parse::<u64>().ok()).unwrap_or(0);
// Convert clock ticks to microseconds (assuming 100 Hz)
// 1 tick = 10ms = 10,000 microseconds
@@ -1167,34 +1289,13 @@ pub async fn collect_process_metrics(
let parent_pid = process.parent().map(|p| p.as_u32());
let start_time = process.start_time();
// Read UID and GID directly from /proc/{pid}/status for accuracy
// Read UID and GID directly from /proc/{pid}/status for accuracy.
// Uses the shared single-read helper (also extracts memory, threads,
// state — we discard those here since sysinfo already provided them).
#[cfg(target_os = "linux")]
let (user_id, group_id) =
if let Ok(status_content) = std::fs::read_to_string(format!("/proc/{pid}/status")) {
let mut uid = 0u32;
let mut gid = 0u32;
for line in status_content.lines() {
if let Some(value) = line.strip_prefix("Uid:") {
// Uid line format: "Uid: 1000 1000 1000 1000" (real, effective, saved, filesystem)
// We want the real UID (first value)
if let Some(uid_str) = value.split_whitespace().next() {
uid = uid_str.parse().unwrap_or(0);
}
} else if let Some(value) = line.strip_prefix("Gid:") {
// Gid line format: "Gid: 1000 1000 1000 1000" (real, effective, saved, filesystem)
// We want the real GID (first value)
if let Some(gid_str) = value.split_whitespace().next() {
gid = gid_str.parse().unwrap_or(0);
}
}
}
(uid, gid)
} else {
// Fallback if /proc read fails (permission issue)
(0, 0)
};
let (user_id, group_id) = read_proc_status(pid)
.map(|s| (s.uid, s.gid))
.unwrap_or((0, 0));
#[cfg(not(target_os = "linux"))]
let (user_id, group_id) = (0, 0);
@@ -1248,6 +1349,9 @@ pub async fn collect_process_metrics(
// Collect thread information (Linux only)
let threads = collect_thread_info(pid);
// One read of /proc/{pid}/stat covers both user + system CPU times.
let (cpu_time_user, cpu_time_system) = get_cpu_times_ms(pid);
// Now construct the detailed info without holding the lock
let detailed_info = DetailedProcessInfo {
pid,
@@ -1264,8 +1368,8 @@ pub async fn collect_process_metrics(
user_id,
group_id,
start_time,
cpu_time_user: get_cpu_time_user(pid),
cpu_time_system: get_cpu_time_system(pid),
cpu_time_user,
cpu_time_system,
read_bytes,
write_bytes,
working_directory,
+4
View File
@@ -17,6 +17,10 @@ pub type SharedNetworks = Arc<Mutex<Networks>>;
pub struct ProcCpuTracker {
pub last_total: u64,
pub last_per_pid: HashMap<u32, u64>,
/// PID → process name cache. Mirrors the non-Linux `ProcessCache.names`.
/// On a Pi with ~150-300 mostly-stable processes this avoids re-allocating
/// the same `String`s on every processes poll (~once per 1.5s).
pub names: HashMap<u32, String>,
}
#[cfg(not(target_os = "linux"))]
+28 -19
View File
@@ -69,12 +69,12 @@ async fn handle_socket(mut socket: WebSocket, state: AppState) {
Message::Text(ref text) if text == "get_processes" => {
let payload = collect_processes_all(&state).await;
// Map to protobuf message
// Get cached buffers
// Get cached buffers. The Vec capacity is preserved across
// calls (with_capacity(512) seeds it, then we swap-back after
// encode so the allocation outlives any single request).
let cache = COMPRESSION_CACHE.get_or_init(|| Mutex::new(CompressionCache::new()));
let mut cache = cache.lock().await;
// Reuse process vector to build the list
cache.processes_vec.clear();
cache
.processes_vec
@@ -85,29 +85,38 @@ async fn handle_socket(mut socket: WebSocket, state: AppState) {
mem_bytes: p.mem_bytes,
}));
let pb = pb::Processes {
// Move the populated Vec into the proto, encode, then move it
// BACK into the cache so the next call reuses the same heap
// allocation. The previous code did `mem::take(...)` here but
// then dropped `pb` (and the Vec along with it), leaving the
// cache holding an empty zero-capacity Vec — defeating the
// whole point of `with_capacity(512)`.
let mut pb = pb::Processes {
process_count: payload.process_count as u64,
rows: std::mem::take(&mut cache.processes_vec),
};
let mut buf = Vec::with_capacity(8 * 1024);
if prost::Message::encode(&pb, &mut buf).is_err() {
let encode_result = prost::Message::encode(&pb, &mut buf);
// Restore the (now-encoded-from) Vec to the cache before pb is
// dropped. We `take` it out of pb to leave that field empty,
// and the next request will `.clear()` before refilling.
cache.processes_vec = std::mem::take(&mut pb.rows);
if encode_result.is_err() {
let _ = socket.send(Message::Close(None)).await;
} else if buf.len() <= COMPRESSION_THRESHOLD {
let _ = socket.send(Message::Binary(buf)).await;
} else {
// compress if large
if buf.len() <= COMPRESSION_THRESHOLD {
let _ = socket.send(Message::Binary(buf)).await;
} else {
// Create a new encoder for each message to ensure proper gzip headers
let mut encoder =
GzEncoder::new(Vec::with_capacity(buf.len()), Compression::fast());
match encoder.write_all(&buf).and_then(|_| encoder.finish()) {
Ok(compressed) => {
let _ = socket.send(Message::Binary(compressed)).await;
}
Err(_) => {
let _ = socket.send(Message::Binary(buf)).await;
}
// Create a new encoder for each message to ensure proper gzip headers
let mut encoder =
GzEncoder::new(Vec::with_capacity(buf.len()), Compression::fast());
match encoder.write_all(&buf).and_then(|_| encoder.finish()) {
Ok(compressed) => {
let _ = socket.send(Message::Binary(compressed)).await;
}
Err(_) => {
let _ = socket.send(Message::Binary(buf)).await;
}
}
}
+1 -2
View File
@@ -1,4 +1,3 @@
use assert_cmd::prelude::*;
use std::fs;
use std::path::PathBuf;
use std::process::Command;
@@ -17,7 +16,7 @@ fn generates_self_signed_cert_and_key_in_xdg_path() {
let xdg = tmpdir.path().to_path_buf();
// Run the agent once with --enableSSL, short timeout so it exits quickly when killed
let mut cmd = Command::cargo_bin("socktop_agent").expect("binary exists");
let mut cmd = Command::new(assert_cmd::cargo::cargo_bin!("socktop_agent"));
// Bind to an ephemeral port (-p 0) to avoid conflicts/flakes
cmd.env("XDG_CONFIG_HOME", &xdg)
.arg("--enableSSL")
+2 -2
View File
@@ -37,9 +37,9 @@ checksum = "46c5e41b57b8bba42a04676d81cb89e9ee8e859a1a66f80a5a72e1cb76b34d43"
[[package]]
name = "bytes"
version = "1.10.1"
version = "1.11.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d71b6127be86fdcfddb610f7182ac57211d4b18a3e9c82eb2d17662f2227ad6a"
checksum = "1e748733b7cbc798e1434b6ac524f0c1ff2ab456fe201501e6497c8417a4fc33"
[[package]]
name = "cfg-if"