Report local Incus health to systemd

On systems running systemd v260 or later, Incus publishes a small set of local health metrics to systemd.

Reporting is enabled automatically when the host’s systemd provides systemd-report.

View the current report with:

systemd-report metrics org.linuxcontainers.Incus

View metric descriptions with:

systemd-report describe org.linuxcontainers.Incus

The systemd report is independent of the /1.0/metrics endpoint and its network configuration. Unlike the Prometheus metrics endpoint, it reports a small snapshot of the operational health of Incus rather than per-instance resource usage.

The following metrics are reported:

  • databaseReadAvailable Is 1 when Incus can successfully perform a read-only database query and 0 otherwise. This does not verify database writes.

  • instanceTypeAvailable Reports the result of the cached Incus support probe for each instance type. It does not attempt to start an instance.

  • unresolvedWarnings Reports the number of new and acknowledged warnings associated with the local node. It counts warning records rather than recurrence counts. Resolved and cluster-wide warnings are excluded.

    If the database cannot be read, this metric is omitted rather than reported as 0.

Systemd itself reports whether the Incus service is running. Incus therefore does not attempt to report whether its own daemon is stopped or unresponsive.

To investigate reported warnings, use:

incus warning list --all

Also check the Incus daemon logs for additional context.

The report socket is available at:

/run/systemd/report/org.linuxcontainers.Incus

It is accessible to local users.