A lightweight monitoring dashboard for VPS servers running containerised apps with Podman. Designed for servers managed by the perc CLI.
- Real-time system metrics (CPU, memory, disk, network)
- Per-container resource usage for Podman workloads
- Historical data stored locally with redb
- Single binary, no external dependencies
- Live-updating dashboard via SSE
Download the latest binary from Releases:
curl -L https://github.com/crabby-utils/perc-stats/releases/download/v0.1.0/perc-stats -o /usr/local/bin/perc-stats
chmod +x /usr/local/bin/perc-statsOr build from source:
cargo install --path .
perc-stats [OPTIONS]
| Flag | Default | Description |
|---|---|---|
--port |
9880 (or PORT env) |
HTTP listen port |
--bind |
0.0.0.0 |
Bind address |
--data-dir |
/var/lib/perc-stats |
Directory for metrics database |
By default no authentication is required, which is fine if you restrict access at the network level (e.g. Tailscale).
To enable HTTP Basic Auth, set one of the following environment variables:
# Plaintext password (hashed with bcrypt at startup)
PERC_STATS_PASSWORD=mysecret
# Or provide a pre-hashed bcrypt password
PERC_STATS_PASSWORD_HASH='$2b$12$...'Optionally set a custom username (defaults to admin):
PERC_STATS_USERNAME=darrenGenerate a bcrypt hash with:
htpasswd -nbBC 12 "" 'yourpassword' | cut -d: -f2The /healthz endpoint is always unauthenticated.
| Path | Description |
|---|---|
/ |
Dashboard UI |
/api/system?range=1h |
System metrics JSON |
/api/containers?range=1h |
Container metrics JSON |
/api/live |
SSE stream of live snapshots |
/healthz |
Health check |
MIT