sensu / sensu-go

Simple. Scalable. Multi-cloud monitoring.
https://sensu.io
MIT License
1.02k stars 176 forks source link

Disable warning state for Keepalive #4235

Open andrew-pickin-epi opened 3 years ago

andrew-pickin-epi commented 3 years ago

I would like the ability to have keep alive go from Ok to Critical without passing through Warning.

Feature Suggestion

If keepalive-critical-timeout is set > 0 it should be possible to disable the intermediate warning state.

Possible Implementation

keepalive-warning-timeout = 0 disabled warning state.

Context

I am using Sensu in concert with Prometheus and Alertmanager.

On the whole this works very well. However Alertmanger model is to have clients post is alert repeatedly, when these stop/expire the alert is considered resolved. Alertmanager sees keepalive warnings and criticals as separate alerts. Thus when the keepalive goes from Warning to Critical, Alertmanager no longer receives Warning posts, and send a Resolved notification (of the warning) even though the entity is still critically down.

mcatngena commented 3 years ago

I also vote for this option.

calebhailey commented 3 years ago

Related: https://github.com/sensu/sensu-go/issues/2883

We should make keepalive and TTL status configurable.