You've already forked alertmanager
mirror of
https://github.com/prometheus/alertmanager.git
synced 2026-10-23 08:48:56 +03:00
b7edc13cfd
MutedByTimeIntervals stopped after the active time intervals muted a group, mirroring the pipeline where the TimeMuteStage never runs after the TimeActiveStage drops the alerts. For a query that only explains why a group is muted, hiding the mute time intervals loses information for no benefit. Evaluate both kinds and return one slice per kind: the active time intervals when now is outside all of them, and the mute time intervals that contain now. The group is still muted exactly when the pipeline would mute it. Carry both on dispatch.AlertGroup as MutedByActiveTimeIntervals and MutedByMuteTimeIntervals so that a future API version can report them separately. The v2 API, whose mutedBy field is a single list, flattens them through the new AlertGroup.MutedBy method, active names first. That method concatenates with slices.Concat because the active names alias the route's own slice. Add a unit test for the combinations, cover the flattening, and update the dispatcher and pipeline tests that pinned the old precedence. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: Guido Trotter <guido@hudson-trading.com>