Evidence Map¶
Use this page to decide which commands and signals matter for each AKS symptom category. Better evidence collection prevents random corrective changes.
Main Content¶
flowchart TD
A[Symptom Category] --> B[kubectl state]
A --> C[Events]
A --> D[Node signals]
A --> E[Azure resource state]
A --> F[Logs and metrics]| Symptom | First Evidence | Follow-up Evidence |
|---|---|---|
| Pod not starting | kubectl get pods, kubectl describe pod | container logs, image pull secret, node conditions |
| Service unreachable | kubectl get svc, kubectl get endpoints | selector labels, NetworkPolicy, DNS lookup |
| Ingress broken | kubectl get ingress, controller logs | load balancer state, backend endpoints, cert/TLS |
| Node unhealthy | kubectl get nodes, kubectl describe node | daemonsets, CNI state, quota, subnet IPs |
| Upgrade issue | cluster version, events | controller compatibility, PDBs, node image rollout |
High-value command groups¶
kubectl get events -A --sort-by=.lastTimestamp
kubectl get pods -A -o wide
kubectl get nodes -o wide
az aks show --resource-group $RG --name $CLUSTER_NAME --output json