[
https://issues.apache.org/jira/browse/SOLR-15767?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17700055#comment-17700055
]
Jan Høydahl commented on SOLR-15767:
------------------------------------
I was just looking into this myself. What remains to get this in?
I can test these out in a test cluster of mine and give some feedback.
Also, it seems as if the prometheus operator installs a bunch of alerts by
default. Should we in solr-operator have a simple way of installing a set of
"safe", universal rules? Perhaps not the "subjective" QPS rules, but the set of
rules that clearly indicate "yellow" or "red" cluster state.
> Include a core set of Prometheus alert rules for monitoring SolrCloud on K8s
> ----------------------------------------------------------------------------
>
> Key: SOLR-15767
> URL: https://issues.apache.org/jira/browse/SOLR-15767
> Project: Solr
> Issue Type: New Feature
> Components: contrib - prometheus-exporter
> Reporter: Timothy Potter
> Assignee: Timothy Potter
> Priority: Major
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> As promised in my ApacheCon talk (recording here:
> https://www.youtube.com/watch?v=mR_3bbqZSRg), I'd like to ship a core set of
> alerting rules for monitoring SolrCloud clusters specifically on Kubernetes.
> The basic guidance will be for users to import these rules into the
> Prometheus stack (which provides a {{PrometheusRule}} CRD). Users will need
> to adjust the thresholds for their specific uses (e.g. p95) and alert
> interval. Eventually, I'd like to have a starter runbook to accompany the
> alert rules, but not there yet.
> Also, it's important to realize that the Prometheus stack provides a core set
> of alerting rules for monitoring node and pod health, so the Solr specific
> set should be considered complementary to those and users will be expected to
> use both sets, see:
> https://github.com/prometheus-operator/kube-prometheus/blob/0821adabf6f9f7aebf9343ecb07707826ce693ee/manifests/kubernetes-prometheusRule.yaml
> If users are not running in Kubernetes or not using the Prometheus stack, the
> rules should still be useful as a getting started guide on what to monitor.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
