cm/cm-dashboard
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 259 Wiki Activity

Reduce all command timeouts to 2-3 seconds max

Browse Source 
With 10-second host heartbeat timeout, all command timeouts must be significantly lower to ensure total collection time stays under 10 seconds.

Changed timeouts:
- smartctl: 10s → 3s (critical: multiple drives queried sequentially)
- du: 5s → 2s
- lsblk: 5s → 2s
- systemctl list commands: 5s → 3s
- systemctl show/is-active: 3s → 2s
- docker commands: 5s → 3s
- df, ip commands: 3s → 2s

Total worst-case collection time now capped at more reasonable levels, preventing false host offline alerts from blocking operations.
This commit is contained in:
Christoffer Martinsson 2025-11-27 16:38:54 +01:00
parent 76c04633b5
commit 374b126446
5 changed files with 26 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
Cargo.lock generated
View File

@ -279,7 +279,7 @@ checksum = "a1d728cc89cf3aee9ff92b05e62b19ee65a02b5702cff7d5a377e32c6ae29d8d"
[[package]]
name = "cm-dashboard"
version = "0.1.186"
version = "0.1.187"
dependencies = [
"anyhow",
"chrono",
@ -301,7 +301,7 @@ dependencies = [
[[package]]
name = "cm-dashboard-agent"
version = "0.1.186"
version = "0.1.187"
dependencies = [
"anyhow",
"async-trait",
@ -324,7 +324,7 @@ dependencies = [
[[package]]
name = "cm-dashboard-shared"
version = "0.1.186"
version = "0.1.187"
dependencies = [
"chrono",
"serde",

6
agent/src/collectors/disk.rs
View File

@ -117,7 +117,7 @@ impl DiskCollector {
let mut cmd = Command::new("lsblk");
cmd.args(&["-rn", "-o", "NAME,MOUNTPOINT"]);
let output = run_command_with_timeout(cmd, 5).await
let output = run_command_with_timeout(cmd, 2).await
.map_err(|e| CollectorError::SystemRead {
path: "block devices".to_string(),
error: e.to_string(),
@ -427,7 +427,7 @@ impl DiskCollector {
cmd.args(&["-a", &format!("/dev/{}", drive_name)]);
}
let output = run_command_with_timeout(cmd, 10).await
let output = run_command_with_timeout(cmd, 3).await
.map_err(|e| CollectorError::SystemRead {
path: format!("SMART data for {}", drive_name),
error: e.to_string(),
@ -764,7 +764,7 @@ impl DiskCollector {
fn get_drive_info_for_path(&self, path: &str) -> anyhow::Result<PoolDrive> {
// Use lsblk to find the backing device with timeout
let output = Command::new("timeout")
.args(&["5", "lsblk", "-rn", "-o", "NAME,MOUNTPOINT"])
.args(&["2", "lsblk", "-rn", "-o", "NAME,MOUNTPOINT"])
.output()
.map_err(|e| anyhow::anyhow!("Failed to run lsblk: {}", e))?;

4
agent/src/collectors/memory.rs
View File

@ -105,8 +105,8 @@ impl MemoryCollector {
return Ok(());
}
// Get usage data for all tmpfs mounts at once using df (with 3 second timeout)
let mut df_args = vec!["3", "df", "--output=target,size,used", "--block-size=1"];
// Get usage data for all tmpfs mounts at once using df (with 2 second timeout)
let mut df_args = vec!["2", "df", "--output=target,size,used", "--block-size=1"];
df_args.extend(tmpfs_mounts.iter().map(|s| s.as_str()));
let df_output = std::process::Command::new("timeout")

2
agent/src/collectors/network.rs
View File

@ -110,7 +110,7 @@ impl NetworkCollector {
// Parse VLAN configuration
let vlan_map = Self::parse_vlan_config();
match Command::new("timeout").args(["3", "ip", "-j", "addr"]).output() {
match Command::new("timeout").args(["2", "ip", "-j", "addr"]).output() {
Ok(output) if output.status.success() => {
let json_str = String::from_utf8_lossy(&output.stdout);

34
agent/src/collectors/systemd.rs
View File

@ -251,18 +251,18 @@ impl SystemdCollector {
/// Auto-discover interesting services to monitor
fn discover_services_internal(&self) -> Result<(Vec<String>, std::collections::HashMap<String, ServiceStatusInfo>)> {
// First: Get all service unit files (with 5 second timeout)
// First: Get all service unit files (with 3 second timeout)
let unit_files_output = Command::new("timeout")
.args(&["5", "systemctl", "list-unit-files", "--type=service", "--no-pager", "--plain"])
.args(&["3", "systemctl", "list-unit-files", "--type=service", "--no-pager", "--plain"])
.output()?;
if !unit_files_output.status.success() {
return Err(anyhow::anyhow!("systemctl list-unit-files command failed"));
}
// Second: Get runtime status of all units (with 5 second timeout)
// Second: Get runtime status of all units (with 3 second timeout)
let units_status_output = Command::new("timeout")
.args(&["5", "systemctl", "list-units", "--type=service", "--all", "--no-pager", "--plain"])
.args(&["3", "systemctl", "list-units", "--type=service", "--all", "--no-pager", "--plain"])
.output()?;
if !units_status_output.status.success() {
@ -358,16 +358,16 @@ impl SystemdCollector {
}
}
// Fallback to systemctl if not in cache (with 3 second timeout)
// Fallback to systemctl if not in cache (with 2 second timeout)
let output = Command::new("timeout")
.args(&["3", "systemctl", "is-active", &format!("{}.service", service)])
.args(&["2", "systemctl", "is-active", &format!("{}.service", service)])
.output()?;
let active_status = String::from_utf8(output.stdout)?.trim().to_string();
// Get more detailed info (with 3 second timeout)
// Get more detailed info (with 2 second timeout)
let output = Command::new("timeout")
.args(&["3", "systemctl", "show", &format!("{}.service", service), "--property=LoadState,ActiveState,SubState"])
.args(&["2", "systemctl", "show", &format!("{}.service", service), "--property=LoadState,ActiveState,SubState"])
.output()?;
let detailed_info = String::from_utf8(output.stdout)?;
@ -427,9 +427,9 @@ impl SystemdCollector {
return Ok(0.0);
}
// No configured path - try to get WorkingDirectory from systemctl (with 3 second timeout)
// No configured path - try to get WorkingDirectory from systemctl (with 2 second timeout)
let output = Command::new("timeout")
.args(&["3", "systemctl", "show", &format!("{}.service", service_name), "--property=WorkingDirectory"])
.args(&["2", "systemctl", "show", &format!("{}.service", service_name), "--property=WorkingDirectory"])
.output()
.map_err(|e| CollectorError::SystemRead {
path: format!("WorkingDirectory for {}", service_name),
@ -449,15 +449,15 @@ impl SystemdCollector {
Ok(0.0)
}
/// Get size of a directory in GB (with 5 second timeout)
/// Get size of a directory in GB (with 2 second timeout)
async fn get_directory_size(&self, path: &str) -> Option<f32> {
use super::run_command_with_timeout;
// Use -s (summary) and --apparent-size for speed, 5 second timeout
// Use -s (summary) and --apparent-size for speed, 2 second timeout
let mut cmd = Command::new("sudo");
cmd.args(&["du", "-s", "--apparent-size", "--block-size=1", path]);
let output = run_command_with_timeout(cmd, 5).await.ok()?;
let output = run_command_with_timeout(cmd, 2).await.ok()?;
if !output.status.success() {
// Log permission errors for debugging but don't spam logs
@ -783,9 +783,9 @@ impl SystemdCollector {
let mut containers = Vec::new();
// Check if docker is available (cm-agent user is in docker group)
// Use -a to show ALL containers (running and stopped) with 5 second timeout
// Use -a to show ALL containers (running and stopped) with 3 second timeout
let output = Command::new("timeout")
.args(&["5", "docker", "ps", "-a", "--format", "{{.Names}},{{.Status}}"])
.args(&["3", "docker", "ps", "-a", "--format", "{{.Names}},{{.Status}}"])
.output();
let output = match output {
@ -826,9 +826,9 @@ impl SystemdCollector {
/// Get docker images as sub-services
fn get_docker_images(&self) -> Vec<(String, String, String, f32)> {
let mut images = Vec::new();
// Check if docker is available (cm-agent user is in docker group) with 5 second timeout
// Check if docker is available (cm-agent user is in docker group) with 3 second timeout
let output = Command::new("timeout")
.args(&["5", "docker", "images", "--format", "{{.Repository}}:{{.Tag}},{{.Size}}"])
.args(&["3", "docker", "images", "--format", "{{.Repository}}:{{.Tag}},{{.Size}}"])
.output();
let output = match output {