Closed jeffmccune closed 7 months ago
It was refreshingly straight forward to use the crunchy monitoring example and add the monitoring key to the postgrescluster spec.
Got to the dashboard using k port-forward svc/crunchy-grafana 3000
but it should be straight forward to add a VirtualService and modify the grafana.ini
like we do.
Some ideas:
Does it make sense to have an authenticated /holos/grafana
path for the environment's host?
How hard is it to configure slack notifications?
Alerts look good, there's a warning for PGDATA at 75% and critical at 90%. Need to test them though.
Closing this for now, we have monitoring merged in. We can revisit how to expose this via the auth proxy in a later ticket.
In addition to a central obs stack, it might be worthwhile to configure a smaller stack for each stage of each project. Particularly for the IAM login service. To start, focus on the database itself with ZITADEL as the use case.
Crunchy has a nice solution for this:
Refer to the Crunchy Postgres for Kubernetes Monitoring stack.