Monitoring Stack was yesterday – Observability Platform at scale!
summary
A B2B Kubernetes fleet (about 150 clusters, 6,000 nodes) moves from brittle monitoring to a platform: core stack is Grafana, Loki, Alertmanager, tracing with Tempo, and a migration Prometheus → Mimir plus OpenTelemetry ingestion. Team reframes monitoring as a self‑service observability platform with a unified entry, clear service vs instance ownership, and three pillars: data, dashboards, alerts. Result: faster feature rollout (tracing deployed in a week), fewer firefights, measurable user adoption and clearer roadmaps.
AI-written summary · about
chapters (10)
- 0:00Introduction and motivation
- 3:27Firefighting mode and challenges
- 7:01Regaining control with product thinking
- 10:37Organizational issues at scale
- 14:15Defining observability platform
- 18:12Platform mindset and target groups
- 21:03Three pillars and self-service
- 22:50Roadmap and platform foundation
- 25:22Product thinking and user experience
- 26:47Wrap-up and key takeaways
related talks
34mThe Grafana Stack: Building a Highly Available, Scalable Observability Platform
29mParanoid Observability –Success Stories and Misfortunes observing an observability stack
27mObservability with Prometheus and beyond
24mPractical Kubernetes Monitoring with Prometheus
36mObservability at Scale With Monitoring as Code...
33mBeyond Monitoring: The Rise of Observability Platform
28mImprove Monitoring and Observability for Kubernetes with OSS Tools
27mMonitoring as Code: Managing your Observability Assets the Cloud-Native Way