Patroni cause alerts for all patroni services
The following discussion from !6568 (merged) should be addressed:
-
@reprazent started a discussion: It's too bad that we've hardcoded this selector here, we now have more than one patroni service that we might to have these alerts for. Would you mind creating an issue crosslinking this and raising it with the database team?
Context
While we're trying to move the alerting rule autogenerated-patroni-cause-alerts.yml
from Prometheus to Thanos, we realized some cause-based alerts with type="patroni"
hardcoded as below:
PostgreSQL_HotSpotTupleFetchingPrimary
PostgreSQL_HotSpotTupleFetchingReplicas
-
PostgreSQLAccessGroupTupleFetchesWarningTrigger
(this is fine because it only checked fromproject_authorizations
table) PatroniLongRunningTransactionDetected
Objective
-
This issue is to discuss whether we want to have these alerts for other patroni services too, i.e.
patroni-registry
,patroni-embedding
,patroni-ci
. -
(For Scalability to decide) If we do want to include these services, we'll also need to decide whether we want scope by the
type
selector too (separate rules bt type), or if we can just query the services in one rule (ietype=~"patroni|patroni-registry|patroni-embedding|patroni-ci"
)