Fool-Proof Kubernetes Dashboards for Sleep-Deprived Oncalls
summary
Shows how to structure Grafana dashboards for Kubernetes so oncall can eliminate whole areas of problems and find the single failing service or instance. Uses Prometheus queries, Kubernetes monitoring mixins, and template-variable driven views to avoid one-off dashboards and 1,800+ sprawl. Demonstrates hierarchical drill-downs (cluster → namespace → pod), normalizing metrics like CPU per core and generating dashboards from higher-order functions stored in GitHub for consistent, small change sets.
AI-written summary · about
chapters (15)
- 0:00Intro
- 2:35Dashboarding for Kubernetes on-calls
- 3:54On-call for Kubernetes
- 4:47The path to 1,000 dashboards
- 5:36Introducing DMM: Dashboarding Maturity Model
- 6:05Dashboarding maturity levels
- 10:10Medium maturity: Methodical dashboards
- 11:48Medium maturity: Hierarchical dashboards
- 13:32Medium maturity: Service hierarchies
- 15:43Medium maturity: Expressive charts
- 17:52Medium maturity: Directed browsing
- 18:33Medium maturity: Managing dashboards
- 20:35High maturity: Optimizing use
- 21:43High maturity: Consistency by design
- 24:08Future workflow: Dashboard as code
related talks
30mCrash course in Kubernetes monitoring
24mPractical Kubernetes Monitoring with Prometheus
19mKeynote: Monitoring Kubernetes Clusters with Prometheus
37mMonitoring Kubernetes with Prometheus
46mHow many dashboards does your Grafana have?
35mThriving With Kubernetes On-Call: Best Practices & Lessons Learned
27mFrom Zero Visibility to Full Observability with Grafana
14mFrom Metrics to Mastery: Monitoring vital signs of Kubernetes