[ https://issues.jboss.org/browse/TEIIDSB-23?page=com.atlassian.jira.plugin... ] Steven Hawkins commented on TEIIDSB-23: --------------------------------------- The 3 scale team is promoting a prometheus instance per openshift cluster, then federating those instances to a master aggregator. I'm not clear yet on the ramifications of having a proliferation of instances - one per namespace, or even one per deployment type... The 30 day retention is also an active topic. Eventually a time-series db will be included in the mix for longer term storage (the initial target is a year), but the may need to be some aggregation/retention rules. We need to validate that our pods at host:8778/metrics are exposing the metrics https://github.com/teiid/teiid-openshift-examples/blob/0a4383e7c9c3514532... Based upon that we can change the https://github.com/teiid/teiid-openshift-examples/blob/0a4383e7c9c3514532... relabel line. The next thing will be some grafana dashboard. The general approach it seems is to keep things as focused as possible (3-4 metrics per board), and a small number of metrics overall - as too much can overwhelm the 30 day retention schedule. I'm not sure yet about the expectation for grafana instances or the delivery for dashboard definitions. -- This message was sent by Atlassian Jira (v7.12.1#712002)