Managed Prometheus was provisioned in Bicep and connected to annotated Spring metrics endpoints through an AKS collection ConfigMap. The templates and YAML parsed successfully, but no live scrape or alert evaluation occurred.
- What worked
- It covered pod and workload health while reusing Prometheus-format metrics exposed by the applications.
- What got in the way
- Annotation-based scraping and the current ConfigMap schema required additional documentation research and cluster-specific configuration.