pm2-dash 0.1.0

A readable web dashboard for pm2 — services, memory, restarts, and live log tailing in one page.

pm2-dash

A readable web dashboard for pm2. Services, memory, CPU, restart counts, and live log tailing in one page. Auto-refreshing, no setup.

I run 57 pm2 services on one box. pm2 monit is a terminal grid that shows a handful at a time and pm2 logs floods every service at once. I wanted the one view pm2 doesn't ship: every service in a sortable table, restart counts I can scan for the one that's flapping, and click-to-tail logs without grepping for a log path. So I built it.

Install

pip install pm2-dash

Run

pm2-dash                 # http://127.0.0.1:9777
pm2-dash --port 8080
pm2-dash --host 0.0.0.0  # expose on the network (put auth in front of it)

It shells out to pm2 jlist for state and reads each service's log files straight off disk. No pm2 stream stays open, and tailing a multi-GB log reads only the tail, not the whole file.

What it shows

• Every service: id, status, CPU, memory, restart count, uptime, exec mode, pid
• Errored and stopped services surfaced in the header; unstable restarts flagged with a ⚠
• Sorted so anything not online floats to the top
• Click any row to tail its stdout + stderr (last 80 lines)
• The whole thing repolls every 2 seconds

Honest scope

Read-only. It does not start, stop, or restart anything, by design, so it's safe to leave open on a shared box. It reads what pm2 reports; if pm2 can't see a process, neither can this. One CDN script tag (htmx) for the polling; no build step, no node_modules. View source is the entire frontend.

If you bind to 0.0.0.0, put a reverse proxy with auth in front of it. It ships no auth of its own.

License

MIT