[
https://issues.apache.org/jira/browse/SOLR-15397?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17340637#comment-17340637
]
Jan Høydahl commented on SOLR-15397:
------------------------------------
While testing I discover that with a 3 node ZK ensemble, if you stop 2 of them,
the /admin/zookeeper/status API will return an exception instead of details
about the Zookeepers. That is since it tries to talk to ZK to get the
/zookeeper/configuration node. This defeats the purpose somewhat, so I intend
to catch that exception and continue talking to each zookeeper over TCP using
ZK_HOST. Then at least this endpoint will always return a JSON response, so you
can know - both in Solr UI and in Grafana, how many Zookeepers you have live
and which ones are dead.
> Expose zookeeper status in the exporter
> ---------------------------------------
>
> Key: SOLR-15397
> URL: https://issues.apache.org/jira/browse/SOLR-15397
> Project: Solr
> Issue Type: New Feature
> Security Level: Public(Default Security Level. Issues are Public)
> Components: contrib - prometheus-exporter
> Reporter: Jan Høydahl
> Assignee: Jan Høydahl
> Priority: Major
> Attachments: Skjermbilde 2021-05-06 kl. 16.48.37.png, Skjermbilde
> 2021-05-06 kl. 16.58.42.png
>
>
> This issue will add these prometheus metrics to solr exporter:
> {{solr_zookeeper_ensemble_size}}, {{solr_zookeeper_nodestatus}},
> {{solr_zookeeper_status}}
> {code:java}
> # HELP solr_zookeeper_nodestatus See following URL:
> https://solr.apache.org/guide/cloud-screens.html#zk-status-view
> # TYPE solr_zookeeper_nodestatus gauge
> solr_zookeeper_nodestatus{host="zoo1:2181",zk_host="zoo1:2181,zoo2:2181,zoo3:2181",}
> 1.0
> solr_zookeeper_nodestatus{host="zoo2:2181",zk_host="zoo1:2181,zoo2:2181,zoo3:2181",}
> 1.0
> solr_zookeeper_nodestatus{host="zoo3:2181",zk_host="zoo1:2181,zoo2:2181,zoo3:2181",}
> 0.0
> # HELP solr_zookeeper_status See following URL:
> https://solr.apache.org/guide/cloud-screens.html#zk-status-view
> # TYPE solr_zookeeper_status gauge
> solr_zookeeper_status{status="yellow",zk_host="zoo1:2181,zoo2:2181,zoo3:2181",}
> 0.0
> # HELP solr_zookeeper_ensemble_size See following URL:
> https://solr.apache.org/guide/cloud-screens.html#zk-status-view
> # TYPE solr_zookeeper_ensemble_size gauge
> solr_zookeeper_ensemble_size{mode="ensemble",zk_host="zoo1:2181,zoo2:2181,zoo3:2181",}
> 3.0
> {code}
> The data is fetched and parsed from Solr's /admin/zookeeper/status endpoint.
> Can be used for alerting on earlier ZK versions that do not natively support
> prometheus, or if you don't want to setup separate ZK monitoring. Example
> Grafana dashboard tile:
> !Skjermbilde 2021-05-06 kl. 16.58.42.png!
> I plan to add this to the new solr cluster row next to the solr live-nodes
> graph:
> !Skjermbilde 2021-05-06 kl. 16.48.37.png|width=800!
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org
