
Grafana's Big Tent Monitoring Kubernetes: Why traditional techniques aren't enough
Oct 18, 2024
Vasil Kaftandzhiev from Grafana Labs and Dio Tsoumas, DevOps lead at GWI, dive into the complexities of monitoring Kubernetes. They discuss why traditional monitoring falls short and explore the evolving landscape of cloud observability. The conversation highlights the growing responsibility for developers and the importance of collaborative ownership. They also cover effective resource management strategies, the vital role of 'golden metrics,' and the intricacies of handling service meshes in resource-demanding environments.
AI Snips
Chapters
Transcript
Episode notes
Out-of-the-Box Observability
- Leverage Kubernetes' out-of-the-box observability tools.
- Developers should focus on code, leaving platform teams to handle monitoring.
Cost as a Monitoring Dimension
- Cloud provider billing adds another layer to Kubernetes monitoring.
- Developers now also share responsibility for managing cloud costs.
Cost Attribution Challenges
- Tom Wilkie recalls struggling with cost attribution in pre-Kubernetes EC2 deployments.
- Kubernetes namespaces simplify cost allocation to services and teams.
