Skip to content
netclaw

Data retention

The daemon deletes its own old log files so the logs directory doesn’t grow forever. Only daemon and crash logs qualify, including the numbered segments the daemon log rolls into: session logs and session data are never pruned. The limit is Retention.Logs.Days, 14 days by default. Change it when you want a longer audit trail or a smaller disk footprint.

What gets deleted

Section titled “What gets deleted”
DeletedNever deleted
daemon-yyyy-MM-dd.log, daemon-yyyy-MM-dd.N.log and crash-*.log files in ~/.netclaw/logs/ whose date is more than Retention.Logs.Days days agoSession logs and session data, including everything in subfolders such as logs/sessions/
The newest 3 daemon log files and the newest 3 crash logs, however old. Segments count as files, so three segments of one day can be the three kept
Any other file in logs/, such as notes.log or a compressed daemon-2026-01-09.log.gz. Only the exact names above match, so a zero-padded daemon-2026-01-09.01.log is left alone

The daemon log rolls over when it reaches 10 MiB: the next lines go to daemon-yyyy-MM-dd.1.log, then .2.log, and so on. A new UTC date starts the series again. Retention prunes the segments by the date in their names, like the first file of the day.

The age comes from the date in the file name, in UTC. Today’s log is never old enough to qualify. The newest-three rule means a wrong system clock can’t wipe your whole history.

When it runs

Section titled “When it runs”

The daemon runs the clean-up 30 seconds after it starts, then every 12 hours. A change to Retention.Logs.Days restarts the daemon like any netclaw.json change, so the new limit takes effect on that restart’s first run. Setting 0 keeps the logs forever and the clean-up deletes nothing.

Change the limit

Section titled “Change the limit”
Terminal window
netclaw config retention                 # show the current value
netclaw config retention --logs-days 30  # keep 30 days
netclaw config retention --logs-days 0   # keep forever

The same screen is in netclaw config under Data Retention: type the number and press Enter. To write the key yourself:

{
  "Retention": {
    "Logs": { "Days": 30 }
  }
}

The environment variable NETCLAW_Retention__Logs__Days=30 works too and wins over netclaw.json. netclaw config retention warns when one is set. Environment changes need a manual daemon restart.

FieldTypeDefaultAllowed values
Retention.Logs.Daysinteger140 to 36500. 0 keeps logs forever.

netclaw config rejects anything outside that range and writes nothing. In a hand-edited file, a value that isn’t an integer falls back to 14 days and the daemon logs a warning at startup. The daemon still starts. A negative number keeps everything, like 0.

Confirm it ran

Section titled “Confirm it ran”

When the clean-up deletes something, the daemon log says so:

02:45:44.817 [INF] Netclaw.Daemon.Services.DataRetentionActor: Retention: deleted 7 expired daemon and crash log file(s) older than 14 days.

A run that finds nothing to delete writes nothing. A file it can’t delete (a read-only volume, say) logs Retention: could not delete N expired daemon and crash log file(s); will retry at the next run. and the daemon keeps going. A bad stored value shows up as:

[WRN] Netclaw.Startup: Configuration warning: Retention:Logs:Days value 'soon' is not an integer; using the default of 14 days.

Look in ~/.netclaw/logs/daemon-<date>.log, or docker exec <container> sh -c 'grep Retention ~/.netclaw/logs/daemon-*.log' in Docker.

Retention elsewhere

Section titled “Retention elsewhere”

Completed one-shot reminders are removed after 12 days.

Section titled “Related pages”

Resources

Section titled “Resources”
Petabridge Built by Petabridge
© 2026 Petabridge, LLC · Apache 2.0