nifikop
nifikop copied to clipboard
konpyutaika
Operator Metrics
Is your feature request related to a problem?
We should add a prometheus metrics endpoint, so the nifikop operator can be monitored
Describe the solution you'd like to see
Some custom metrics, similar to the prometheus operator
Things like total reconcile operations, error count etc
This could help spot operator bugs
Describe alternatives you've considered
No response
Additional context
No response
I discovered through working on this that most of the metrics in the linked Prometheus example are already provided by the controller runtime. There may still be additional metrics we'd like to capture but we can specify those in this issue before raising another PR.
https://book.kubebuilder.io/reference/metrics-reference.html
Off the top of my head, we could capture metrics around the following:
- Number, response code, and duration of requests-by-API-endpoint to NiFi nodes/clusters
- Number of CRDs a nifikop pod is actively watching
We can use a similar Metrics setup as i used in #154