Full Observability for Kyverno With Observability-as-Code
summary
A local kind cluster uses Argo CD to reconcile Kyverno, Prometheus, Tempo and Grafana; Kyverno emits OpenTelemetry spans so rule execution shows per-span timing and failure reasons. Policies include an enforcing demo_approved=true label and audit-only team checks; Grafana is auto-provisioned from Git (dashboards, alerts, data sources) and an alert fires when policy failures exceed 2 in 5 minutes (demo shows >0 failures). No OpenTelemetry Collector or SLO dashboard is deployed; tracing successfully pinpoints rule-level latency.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:45Kyverno needs observability
- 3:30GitOps and observability-as-code
- 4:36Traces beyond metrics
- 5:38Demo setup and architecture
- 9:10Repository structure
- 13:23Auto-provisioning dashboards
- 14:05Kyverno dashboards in Grafana
- 15:30Defining alerts
- 16:34Kyverno policy enforcement
- 17:58Tracing policy execution
- 20:23Future improvements
- 21:45Final takeaways and resources
related talks
26mUnifying Policy Observability in Kubernetes
27mUnderstanding Kyverno in a Visual Way
46mKyverno: Compliance Inside Your Kubernetes Cluster
49mKubernetes Policy as Code with Kyverno
06mProject Lightning Talk: Kyverno: Kubernetes Policy as Code Unleashed!
25mGrafana as Code-An Open Source Observability Solution
36mObservability at Scale With Monitoring as Code...
31mKubernetes Native Policy Management with Kyverno!