Logs and status

Show me the last 50 lines from mod-app.

You get recent output as your program printed it. Lines prefixed with [spocket] are ours — the fetch, the install, the start. Everything else is your code.

Logs are a rolling window, not an archive. When the file reaches your plan’s ceiling the oldest half is dropped, so how far back you can see depends on how much your bot prints.

Those durations assume a normally chatty app. An app logging every event will fill its window faster. If you need something kept permanently, write it to your own database.

When a log is long, reading the tail is the wrong tool. Search it instead:

Search mod-bot’s logs for "rate limit".

You get matching lines with their position and a little surrounding context, which is usually enough to see what led to something. The panel console has the same thing as a filter box — it narrows what is displayed while the output keeps updating.

Status reports the container, not a stored guess — so an app that died a second ago says so rather than showing stale good news.

Ask your agent rather than reading logs yourself — it can read them, find the cause, change the code and redeploy in one turn.

mod-app keeps stopping. Read its logs, work out why, fix it and redeploy.

• A missing environment variable — the thing your .env had that the deploy did not.
• A package your code imports but package.json never declared.
• An invalid or reset token. Discord shows a token once; resetting it invalidates the old one.
• An entry point that does not match the file that is actually there.