KlusterAlert

AI Explanations

Every alert on Pro and above includes an AI-generated plain-English explanation of what happened, the most likely root causes, and the exact kubectl command to start investigating, automatically, with no configuration required.

Available on Pro, Team, and Enterprise plans

What's included

When KlusterAlert detects an issue, it calls an AI model with the pod name, issue type, restart count, and cluster context. The response is appended to every notification sent to Teams, Slack, or email. No extra setup is needed. It works out of the box on eligible plans.

What is happening

Plain-English summary of the detected issue, no Kubernetes jargon

Most likely causes

The 2–3 most common root causes for this specific issue type

kubectl command

The exact command to run first: logs, describe, or events

Example: CrashLoopBackOff

AI explanation: CrashLoopBackOff

🔴 CRITICAL · payments-service · namespace: production

What's happening:
Pod 'payments-service' is in CrashLoopBackOff with 7 restarts.
The container exits immediately after starting.

Most likely causes:
- Missing required environment variable or secret
- Application bug causing a fatal error on startup
- Incompatible Docker image tag recently deployed

Investigate with:
kubectl logs payments-service -n production --previous

Example: OOMKilled

AI explanation: OOMKilled

🔴 CRITICAL · worker-service · namespace: jobs

What's happening:
The container was killed by the kernel because it exceeded its memory limit.
This is an OOMKilled event. The container hit the limit set in the pod spec.

Most likely causes:
- Memory limit set too low for the current workload
- Memory leak introduced in a recent deployment
- Sudden increase in request volume or job batch size

Investigate with:
kubectl describe pod worker-service -n jobs

On the Free and Starter plans, alerts include a single-sentence description of the issue type without AI-generated root-cause analysis. Upgrade to Pro to get full AI explanations.

Covered issue types

AI explanations are generated for all detected issue types: CrashLoopBackOff, OOMKilled, Pod Failed, ImagePullBackOff, pod stuck Pending, high memory usage, and CPU spikes. The explanation is tailored to the specific issue and includes context from the cluster name, namespace, and recent restart history.