The liveness probe that makes your uptime worse
Liveness, readiness and startup probes do different jobs. Conflating them is one of the few Kubernetes misconfigurations that actively causes outages.
Tag
Reliability
Every post tagged Reliability — practitioner notes from running Kubernetes and DevOps tooling in production for clients across the EU and the Gulf.
3 posts
Requests and limits — CPU and memory are not analogous
What happens when a container exceeds its CPU limit versus its memory limit, why the two answers differ, and what that means for how you set them.
PodDisruptionBudgets: the reason your node drain hangs
What a PDB does and does not protect against, the single-replica trap that blocks upgrades forever, and how to set minAvailable so maintenance can still happen.