A lack of monitoring on Solr could cause bug 52148 to go unnoticed for a while (about a week).
This issue aims to have Solr monitored in such a way that "unexpected behaviour" would be noticed automatically, and <relevant roles> would be notified, so that service outages would be detected sooner and can be responded to quicker.
Version: wmf-deployment
Severity: enhancement
See Also:
https://bugzilla.wikimedia.org/show_bug.cgi?id=52148
I was thinking whether we can have monitors that check the number of documents in the index, and alert if that drops a lot?
Elasticsearch won over Solr (see https://www.mediawiki.org/wiki/Search#Solr_vs_Elasticsearch ) so this is INVALID now.
Elasticsearch monitoring is covered by bug 57210.
Andre: this is about TTMServer, which is still using Solr.
However, since we will be working migration to ElasticSearch for TTMServer, it makes no point to work on this bug.