Nagstamon icon indicating copy to clipboard operation
Nagstamon copied to clipboard

Alertmanager alerts do not show up if neither WARNING nor CRITICAL

Open gerdriesselmann opened this issue 3 years ago • 8 comments

Nagstamon version: 3.8.0 and earlier

What causes issue ?

We have Prometheus alerts that have a severity of "unreachable". Nagstamon reads this via Alertmanager.

Expected behavior?

The alerts show up as "UNREACHABLE" in Nagstamon and cause an according alert.

This is how I understand the method "map_severity" in the AlertmanagerServer class.

What happens instead?

Alerts are neither shown in status window list of alerts nor does a notification appear.

Any more details?

If I add "unreachable" to the list in the server setting "Map to CRITICAL", the alerts are indeed classified as CRITICAL and do show up.

This IMO shows they have been successfully retrieved and are indeed discarded later on by Nagstamon for some reason.

All filters are unchecked. Notifications are enabled for every status (UNKNOWN to DOWN)

gerdriesselmann avatar Jan 13 '22 16:01 gerdriesselmann

@stearz any idea?

HenriWahl avatar Jan 23 '22 08:01 HenriWahl

I can confirm that this bug is still present in 3.10.1.

Noki avatar Nov 10 '22 10:11 Noki

@gerdriesselmann @Noki did it work with ANY version of Nagstamon before?

HenriWahl avatar Nov 14 '22 10:11 HenriWahl

@HenriWahl I unfortunately don't know. We introduced the UNREACHABLE severity in December 2021 when 3.8 was the latest release.

gerdriesselmann avatar Nov 14 '22 14:11 gerdriesselmann

@gerdriesselmann is this fixed for you with a newer testing release?

HenriWahl avatar Dec 12 '23 06:12 HenriWahl

@HenriWahl Unfortunately not :( I tried 3.13-20240108

gerdriesselmann avatar Jan 18 '24 09:01 gerdriesselmann

@stearz do you have any hint here?

HenriWahl avatar Feb 03 '24 23:02 HenriWahl

I'm afraid this is a limitation in alertmanager as it only supports service objects but the status UNREACHABLE only exists in Nagstamon's Server objects not in the Service objects. So unless there will be a big refactoring of Nagstamon I think the "solution" will be to add the status "unreachable" to the "Map to CRITICAL" like you already did.

stearz avatar Feb 05 '24 09:02 stearz