Your infra runs on K8s, and your controllers and custom resources hum happily. Until they don't. One change later and… UH OH. The logs are a never-ending scroll of problems, dashboards flash like disco lights, and you are knee-deep in debugging hell. You check your custom resources, and they offer no guidance. You eventually fix it, add better error messages, and feel proud. And then it breaks again, with a brand new, even less helpful error.
Why is this so hard?
In this talk, Derik, a prolific controller-writer, and Cat, a regular listener to DevOps meltdowns, will walk you through tested strategies for making your K8s resources less mysterious and more manageable. You will learn how to:
- Add meaningful, traceable errors
- Make events and status your best friends
- Build dashboards that humans want to use
- Set up business-relevant alerts, not just CPU spikes
- Automate away the slog using AI
Make your controller logs and custom resource statuses a source of clarity, not chaos!