Logging

Mitra logs to standard output, so docker compose logs shows everything:

docker compose logs -f mitra

A healthy server is quiet on purpose: by default it only logs what matters. Turn the level up while you track something down, and back down when you’re done.

Set the level

environment:
  MITRA_LOG_LEVEL: 'debug'

Each level includes everything quieter than it:

MITRA_LOG_LEVEL	What you get
error	Failures only.
warn	Also problems Mitra worked around, such as a reminder that couldn’t be delivered, a geocoder that didn’t answer, or a sign-in provider it couldn’t reach.
info (default)	Also what matters day to day: starting up, sign-ins, accounts connected, changes synced from providers and reminders sent.
debug	Also every request with its status and time, each sync, when syncing speeds up or slows down as people open and close Mitra, sessions, entry edits and the conversations with CalDAV servers.
trace	Also every database query and the raw calendar data. There’s a lot of it.

Mitra logs the level it runs at when it starts.

NOTE

Passwords, tokens and other secrets are never logged, at any level. debug and trace can still show entry titles and calendar data, so look through a log before you share it.

Which level to use

• When a calendar isn’t syncing, debug shows each sync and the requests to CalDAV and Notion.
• When reminders don’t arrive, info already logs each reminder as it goes out, and debug adds each delivery attempt and devices that were dropped.
• When you see an error page, error has it with a stack trace, and the default info includes it.
• When you need to see exactly what a provider sent, trace adds the raw data and the database queries. Only leave it on briefly.