Ethan Hunter

Alertmanager service discovery does not update under heavy load

7

### What did you do? We have a cluster of Alertmanager instances running in HA mode. We've configured Prometheus to discover the Alertmanager instances using DNS SRV SD. In our...

Re-add group key to alert groups endpoint

2

I'm working on some integrations which would benefit from associating webhook notifications with alerts returned from the results from `alerts/groups`. It'd be super useful if this endpoint included the group...

bugfix: only increment Silences version if a silence is added

8

This is a small change which adds a second return value to `state.merge` to differentiate between cases where a silence is updated in place and where a silence is added....

Signficantly improve inhibitor performance via new cache datastructure

6

This PR improves or matches performance of the existing inhibitor in almost all benchmark cases. When each inhibition rule has many inhibiting alerts, performance is improved by several orders of...

`Groups` calls can take a long time when there are many aggrGroups or when on of the filter functions is slow. Right now, `Groups` holds the `Dispatcher` lock for the...

This change adds the group key back to the `AlertGroup` struct so that it's available in the `groups` API endpoint. I started a discussion about this with #3446. For the...

Add new behavior to avoid races on config reload

1

When the config is reloaded, alertmanager re-creates the `inhibit.Inhibitor` and the `dispatch.Dispatcher`. This is necessary because both the `Inhibitor` and `Dispatcher` have internal state which depends on the config. In...