jason/socktop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

protobuff Process list

Browse Source 
BREAKING: Process list over WS is now Protocol Buffers; client required.
Agent: returns all processes (no server-side top-k); large payloads gzip-compressed.
Client: decodes protobuf (gz/raw), moves sorting/pagination to TUI.
Build: add prost/prost-build with vendored protoc; enable thin LTO, panic=abort, strip symbols.
Cleanup: cfg-gate Linux-only code; fix Clippy across platforms; tests updated (ws probe TLS CA).
This commit is contained in:
jason
2025-08-19 23:24:36 -07:00
parent 10501168c5
commit 554a2c349f
14 changed files with 327 additions and 144 deletions
Download Patch File Download Diff File Expand all files Collapse all files
socktop_agent/src/metrics.rs
+7 -24
View File
@@ -236,9 +236,9 @@ fn read_proc_jiffies(pid: u32) -> Option<u64> {
Some(utime.saturating_add(stime))
}
/// Collect top processes (Linux variant): compute CPU% via /proc jiffies delta.
/// Collect all processes (Linux): compute CPU% via /proc jiffies delta; sorting moved to client.
#[cfg(target_os = "linux")]
pub async fn collect_processes_top_k(state: &AppState, k: usize) -> ProcessesPayload {
pub async fn collect_processes_all(state: &AppState) -> ProcessesPayload {
// Fresh view to avoid lingering entries and select "no tasks" (no per-thread rows).
let mut sys = System::new();
sys.refresh_processes_specifics(
@@ -291,7 +291,7 @@ pub async fn collect_processes_top_k(state: &AppState, k: usize) -> ProcessesPay
.collect();
return ProcessesPayload {
process_count: total_count,
top_processes: top_k_sorted(procs, k),
top_processes: procs,
};
}
@@ -317,13 +317,13 @@ pub async fn collect_processes_top_k(state: &AppState, k: usize) -> ProcessesPay
ProcessesPayload {
process_count: total_count,
top_processes: top_k_sorted(procs, k),
top_processes: procs,
}
}
/// Collect top processes (non-Linux): use sysinfo's internal CPU% by doing a double refresh.
/// Collect all processes (non-Linux): use sysinfo's internal CPU% by doing a double refresh.
#[cfg(not(target_os = "linux"))]
pub async fn collect_processes_top_k(state: &AppState, k: usize) -> ProcessesPayload {
pub async fn collect_processes_all(state: &AppState) -> ProcessesPayload {
use tokio::time::sleep;
let mut sys = state.sys.lock().await;
@@ -354,8 +354,6 @@ pub async fn collect_processes_top_k(state: &AppState, k: usize) -> ProcessesPay
mem_bytes: p.memory(),
})
.collect();
procs = top_k_sorted(procs, k);
ProcessesPayload {
process_count: total_count,
top_processes: procs,
@@ -363,19 +361,4 @@ pub async fn collect_processes_top_k(state: &AppState, k: usize) -> ProcessesPay
}
// Small helper to select and sort top-k by cpu
fn top_k_sorted(mut v: Vec<ProcessInfo>, k: usize) -> Vec<ProcessInfo> {
if v.len() > k {
v.select_nth_unstable_by(k, |a, b| {
b.cpu_usage
.partial_cmp(&a.cpu_usage)
.unwrap_or(std::cmp::Ordering::Equal)
});
v.truncate(k);
}
v.sort_by(|a, b| {
b.cpu_usage
.partial_cmp(&a.cpu_usage)
.unwrap_or(std::cmp::Ordering::Equal)
});
v
}
// Client now handles sorting/pagination.