Surface failing Kubernetes resources for the current cluster.
This command inspects the live cluster via the Kubernetes API and reports
any pods that are not running successfully, any nodes that are not Ready,
and any PersistentVolumeClaims stuck in Pending phase. Each finding
includes a severity classification and, where a known failure pattern is
detected, a proactive remediation suggestion.
A health score (0–100) summarises overall cluster health: each critical
finding deducts 25 points and each warning deducts 10 points.
The output is intentionally compact so it can be consumed directly by users
or by the KSail AI chat assistant (ksail chat) and MCP server, which expose
this command as part of the cluster_read tool. When used from the AI
assistant the output is fed back as context so Copilot can explain the root
cause and suggest remediation.
The cluster is resolved in the following priority order:
2. From ksail.yaml config file (if present)
3. From current kubeconfig context
Exit code 0 is returned even when pod or node failures are reported.
A non-zero exit code indicates the Kubernetes API could not be queried
(e.g., the cluster is unreachable or the credentials lack sufficient
ksail cluster diagnose [flags]
-n, --name string Name of the cluster to target
--output string Output format: text or json. Use json for machine-readable structured output. (default "text")
-p, --provider Provider Provider to use (Docker, Hetzner, Omni, AWS, Kubernetes)
--benchmark Show per-activity benchmark output
--config string Path to config file (default: ksail.yaml found via directory traversal)
