← Back to guidesGuide

Investigate CrashLoopBackOff with PromQL before blaming OOMKilled

An actionable flow for combining kube-state-metrics, cAdvisor, events, and previous logs without turning a clue into a closed diagnosis.

Learn how to move from a CrashLoopBackOff alert to a verifiable OOMKilled hypothesis with PromQL, kubectl, and validation rules that avoid false pages.

Created: August 8, 2026

Published: August 8, 2026

Estimated time35 min

LevelIntermediate

Before you startPrometheus or a PromQL-compatible platform collecting kube-state-metrics and container metrics.

PlatformsLinux / Docker

WhatsApp X LinkedIn

Linux

Validation from a Linux workstation with kubectl, promtool, and HTTP access to Prometheus.

kubectlpromtoolcurlread-only kubeconfig

Query recent restarts

promtool query instant http://prometheus:9090 'sum by (namespace,pod,container) (increase(kube_pod_container_status_restarts_total[15m]))'

Inspect the affected pod

kubectl describe pod checkout-api-7c9d -n prod

Read the previous container logs

kubectl logs checkout-api-7c9d -n prod --previous --tail=200

Also available inInvestigar CrashLoopBackOff con PromQL antes de culpar a OOMKilled

Content locked

This guide requires both steps before full content is available.

○Click “Like” on this guide.
○Share on WhatsApp, X, LinkedIn, or copy the link.

Access is automatically unlocked as soon as both steps are completed.