Q: Explain how Kubernetes handles pod eviction during node memory pressure.
kubelet monitors node memory usage against eviction thresholds:
Want to master this scenario in a live sandbox? KodeKloud's CKA & CKAD Hands-On Certification Track covers this exact problem with hands-on terminal drills.
🛠️ Production Runbook & Step-by-Step Resolution
Direct Answer: "In our production Kubernetes clusters running microservices on EKS/AKS, this was a classic operational challenge. When addressing this question, I walk the interviewer through our production incident runbook: isolating the blast radius, checking diagnostic logs and metrics, and applying a safe fix."
Initial Diagnostics & Root Cause Analysis
kubelet monitors node memory usage against eviction thresholds:
- Soft eviction — e.g.,
memory.available < 500Mi. kubelet gives pods a grace period (eviction-soft-grace-period, default 90s) to shut down before killing. - Hard eviction — e.g.,
memory.available < 100Mi. kubelet immediately kills pods with no grace period. - BestEffort pods first.
Remediation & Permanent Safeguards
Eviction order: After eviction, the node reports MemoryPressure condition and is tainted. New pods won't schedule there until pressure resolves.
- Burstable pods that exceed their memory request (most over request first).
- Guaranteed pods (last resort).
- Soft eviction — e.g., memory.available . kubelet gives pods a grace period (eviction-soft-grace-p...
- Hard eviction — e.g., memory.available . kubelet immediately kills pods with no grace period.
- BestEffort pods first.
Senior hiring managers and bar raisers evaluate how you prioritize during outages. Avoid these common anti-patterns:
- Restarting a crashing container without examining preceding termination logs via kubectl logs --previous when configuring Explain Kubernetes handles Restarting the pod flushes ephemeral container runtime state, destroying vital panic stack traces and fatal initialization logs.
-
Misdiagnosing an application boot crash as a failure of aggressive liveness probes when configuring Explain Kubernetes handles
Liveness probes with short timeouts or high failure thresholds kill slow-starting applications repeatedly, causing false
CrashLoopBackOffloops. -
Neglecting node taints, tolerations, and resource quotas when diagnosing pods stuck in Pending when configuring Explain Kubernetes handles
Pods remain
Pendingwhen scheduler predicates fail;kubectl describe podreveals whether CPU/memory exhaustion or taints prevent placement. - Invoking kubectl delete pod --force --grace-period=0 indiscriminately on unready pods when configuring Explain Kubernetes handles Forced deletion bypasses kubelet cleanup and can cause multi-attach errors on PersistentVolumes (PVs) or split-brain on stateful workloads.