Rotate your logs before they rotate you
A full disk is the most boring outage there is. Set size limits in journald, give every service a logrotate rule, and check once a quarter that the rules still match the paths the services actually write to.
Short notes on keeping small servers tidy.
A full disk is the most boring outage there is. Set size limits in journald, give every service a logrotate rule, and check once a quarter that the rules still match the paths the services actually write to.
Systemd timers keep their own history, survive missed runs and show up in one list. Moving the nightly jobs took an evening and removed a whole class of “did it run?” questions.
The config already says what is set. A one-line comment about why it is set that way is the part nobody can reconstruct a year later.