Owner: Architect:

Story (Required)

As an OpenShift helm cluster operator, I will like have health state metrics for all helm releases, so I can determine which helm chart are used and whether they are running in healthy state.

Background

We need to start creating metrics for helm at the local cluster level, so that we can later promote them to be aggregated at the Thanos level.

Glossary

Prometheus: https://prometheus.io/docs/prometheus/latest/getting_started/

Out of scope

Not sending metrics to be aggregate in Thanos in this story

In Scope

Local Prometheus metrics

Approach(Required)

We need to create a metrics endpoint and register with Prometheus. The firs metric we will focus on this story will be helm_chart_release_health_status. The value will be either 0 or 1 with 1 being healthy and 0 being unhealthy. There are three dimension/properties attached to the metric:

1. The name of the chart
2. The name of the release
3. The version of the chart

We will need to study how Prometheus work in general and how it is being used to send metrics in redhat. We can study the way samples operator does this here: https://github.com/openshift/cluster-samples-operator/blob/master/pkg/metrics/server.go

Dependencies

Prometheus is deployed and configured in OpenShift cluster

Edge Case

NA

Acceptance Criteria

We can see the metric helm_chart_release_health_status with all it's dimensions for each release deployed
There is a wiki section explaining how to add local metrics to Prometheus

INVEST Checklist

Dependencies identified
Blockers noted and expected delivery timelines set
Design is implementable
Acceptance criteria agreed upon
Story estimated

Legend

Unknown
Verified
Unsatisfied