NodeRingsDocs

Verify and diagnose

Run health checks, collect logs, and produce a debug bundle with nr

After registration — or anytime an agent misbehaves — use the built-in verify and diagnostic commands.

Verify (pass/fail gate)

nr cluster register runs verify at the end. Re-run it explicitly:

nr cluster verify --name edge-ams-01

Limit to subsections:

nr cluster verify --name edge-ams-01 --section kubernetes,calico,liqo,peering

Valid section names: kubernetes, calico, liqo, peering, offloading, agent, operator.

Useful flags:

FlagPurpose
--skip-operatorDo not require operator workloads (same idea as register --skip-operator-install)
--disable-offloadingDo not require NamespaceOffloading checks
--no-agent-apiSkip platform API agent/provisioned checks
--output jsonMachine-readable report on stdout (logs stay on stderr)

Failed verify exits 1. Bad flags / bad --output / bad --section exit 2.

Warn and skip results do not fail the report; only required failures do.

Status, info, health

nr cluster status --name edge-ams-01   # local phase + reachability
nr cluster info --name edge-ams-01     # agent metadata, versions, CIDRs
nr cluster health --name edge-ams-01   # component health checks

Add --output json when scripting.

Logs and debug bundle

nr cluster logs --name edge-ams-01 --lines 200
nr cluster debug --name edge-ams-01

debug writes a report under ~/.nr/debug-<timestamp>/ suitable for support tickets.

Redact secrets before sharing a debug bundle. Remove tokens, kubeconfig credentials, and Proxmox secrets from any files you attach.

Common failure patterns

AreaFirst checks
Authnr auth status; renew NR_API_TOKEN (service account). Confirm --org-id
Tool downloadsNetwork to GitHub Releases; inspect ~/.nr/bin
Ports / prechecks6443 free; enough disk/RAM; Ubuntu
PeeringCorrect --agent-ip and --gateway-region; nr cluster health
OperatorSecrets present; pods in the operator namespace; Proxmox URL is https://

Broader provider troubleshooting: Provider troubleshooting.