KubeMortem is a black box flight recorder for Kubernetes. When pods crash, it captures logs, events, and resource state — so you can debug after the pod is gone.
Works with
The reality
Crashes are inevitable. Losing the forensic evidence shouldn't be.
When a pod enters CrashLoopBackOff or gets OOMKilled, Kubernetes restarts or replaces it. The logs, events, and state that explain why are gone before you even get paged.
You scramble to kubectl logs before the container is gone. You check events but they've been garbage-collected. You reconstruct from memory what the pod spec even looked like.
Crashes happen across namespaces and clusters. Without a central view, recurring patterns go unnoticed and the same issues keep waking your team up at 3 AM.
How it works
Register your Kubernetes clusters and install the KubeMortem agent via a one-liner Helm command. Agents connect outbound — no ingress or stored credentials required.
When a pod enters CrashLoopBackOff, gets OOMKilled, or exits with an error, the agent captures container logs, pod spec, events, and node conditions — before Kubernetes cleans up.
Browse crash events in the dashboard or CLI. Every crash includes the complete forensic snapshot — logs, events, resource state — so you can diagnose even after the pod is long gone.
Features
Built for DevOps teams who are tired of debugging crashes after the evidence is gone.
Automatically detects CrashLoopBackOff, OOMKilled, and error-state pods. Captures logs, events, and resource snapshots before Kubernetes garbage-collects the evidence.
Each crash event stores container logs, pod spec, node conditions, recent events, and resource metrics — everything you need to diagnose after the fact.
See every crash across your fleet in a single chronological timeline. Filter by cluster, namespace, or crash reason to spot patterns fast.
Terminal-first workflow with kubemortem CLI. Visual dashboard for team-wide crash visibility, triage, and historical analysis.
Deploy a lightweight agent via Helm — no public API server or stored credentials needed. Agents watch for crashes and stream forensic data outbound to KubeMortem.
All crash data is encrypted at rest with AES-256-GCM. Your pod logs and resource state never leave your control in plaintext.
Pricing
All plans include crash capture, CLI access, AES-256 encryption, and agent-based collection.
Start capturing crashes for free.
The best value for most teams.
For regulated industries and enterprise.
Sign up in 30 seconds. No credit card required. Start capturing crashes for free.