Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.suji.fr/llms.txt

Use this file to discover all available pages before exploring further.

The Logs tab streams live docker compose logs output from your VM into the dashboard. Open it, pick which install you want, watch.

​
What you see

  • Live stdout + stderr from the install’s containers.
  • Color-coded levels when the app emits them (info / warn / error).
  • Timestamps on each line.
  • A pause toggle if a chatty app is making the view unreadable.
  • Search/filter to grep through what’s on screen.

​
Picking an install

The selector at the top lists every install on this VM. Pick one — the view streams its logs. Switch selectors any time without leaving the page. There’s no “host logs” view; the host itself doesn’t run a single log stream that’s meaningful. Use journalctl from the Terminal tab if you need system-level logs.

​
Scrollback

The view holds the most recent ~10 000 lines. For older logs:
  • From the Terminal tab: docker compose -f /etc/suji/installs/<install-id>/compose.yaml logs --tail=10000 --since=24h <service>
  • Or journalctl --since=... if you want host-side context.
We don’t ship logs to a remote store by default. If you need centralised logging, install a log shipper as a sidecar (Vector, Promtail, etc.) and point it at your collector.

​
Common patterns

Logs tab → pick the install → scroll to the top of the most recent restart. The startup error (database connection refused, EACCES on a data path, missing env var) is almost always in the first 30 lines.
Open the install’s Logs. Trigger the webhook from the third-party service. If you don’t see any inbound activity, the request never reached the app — check the Cloudflare Tunnel and DNS for <install-subdomain>.suji.fr. If you see a 4xx/5xx, the app rejected it; the error body will tell you why.
Logs alone aren’t enough — pair them with the Metrics tab to see CPU/memory pressure at the same time as the slow events.

​
Next

Terminal

Run ad-hoc commands when the logs aren’t enough.

Network

Inspect the public IP, hostname, and traffic counters.