Details
-
Improvement
-
Resolution: Unresolved
-
Major
-
None
-
None
Description
Most alerts are defined with labels that already exist. Whatever labels returned from the expression / PromQL are automatically added to the alert and sent to AlertManager, there is no reason to specify them. For example the following labels can be removed:
- job
- cluster
- bucket
- node
These will automatically be available in the triggered alert, along with the "instance" label and whatever else is returned from the query.
- alert: CB90061-diskWriteQueueLength-Warn
|
expr: |
|
kv_ep_storage_age_seconds > 50 < 100 |
for: 1m |
labels:
|
job: couchbase_prometheus
|
kind: bucket
|
health_check_id: CB90061
|
health_check_name: diskWriteQueueLength
|
cluster: '{{ $labels.cluster }}' |
bucket: '{{ $labels.bucket }}' |
severity: warning
|
annotations:
|
title: "Long disk write queue on Bucket: {{ $labels.bucket }}" |
description: The age of items in the disk write queue is over 50 seconds. |
remediation: Review your hardware for malfunctions or sizing issues. |
If the problem persists, then please contact Couchbase Technical Support.
|
|