Kubernetes pod failure alerts to your phone.

Why Nerve for Kubernetes

Most Kubernetes monitoring stacks give you dashboards and email digests. Nerve gives you an encrypted push alert on your phone the moment a pod dies — from a one-line shell command, not a YAML-heavy operator.

The sender DSN you store in your cluster can only send signals into one pipe. If it leaks, it cannot read your alert history or run commands on your infrastructure.

Watch failed pods with kubectl

The simplest approach: a kubectl one-liner that watches for pod failures and pipes them to Nerve.

kubectl get events --field-selector reason=Failed --watch-only \
  -o custom-columns=MSG:.message --no-headers \
  | while read -r line; do echo "$line" | nerve send; done

CronJob: periodic pod health check

Run a Kubernetes CronJob every 5 minutes that checks for crashing pods and sends a summary.

apiVersion: batch/v1
kind: CronJob
metadata:
  name: nerve-pod-check
spec:
  schedule: "*/5 * * * *"
  jobTemplate:
    spec:
      template:
        spec:
          containers:
          - name: check
            image: bitnami/kubectl:latest
            env:
            - name: NERVE_DSN
              valueFrom:
                secretKeyRef:
                  name: nerve-credentials
                  key: dsn
            command:
            - /bin/sh
            - -c
            - |
              CRASHED=$(kubectl get pods --all-namespaces \
                --field-selector=status.phase=Failed -o name)
              if [ -n "$CRASHED" ]; then
                echo "crashed pods: $CRASHED" | nerve send --severity critical
              fi
          restartPolicy: Never

Store the DSN as a Kubernetes secret

kubectl create secret generic nerve-credentials \
  --from-literal=dsn="nerve://TOKEN:[email protected]"

OOMKill alerts

OOMKilled containers often restart silently. Catch them before they cascade.

kubectl get events --field-selector reason=OOMKilling --watch-only \
  -o custom-columns=MSG:.message --no-headers \
  | while read -r line; do echo "OOMKill: $line" | nerve send --severity critical; done

Deployment rollout failures

kubectl rollout status deployment/my-app --timeout=120s \
  || echo "Rollout failed: my-app" | nerve send --severity critical

FAQ