Troubleshoot Wunderbox safely

Diagnose from consumer outcome toward shared dependencies. Avoid making several unrelated changes at once; each change can hide the original failure and widen the affected failure domain.

Safe triage

1. Protect data integrity and preserve administrative access.
2. Freeze unrelated changes and record the UTC time window.
3. Identify affected consumers, services, and failure domains without copying real identifiers into public channels.
4. Compare consumer-facing health with resource, platform-service, dependency, and management-layer health.
5. Check recent approved changes and immutable artifact identities.
6. Determine whether recovery objectives or capacity thresholds are at risk.
7. Apply one reversible, authorized diagnostic or recovery action at a time.
8. Verify both platform state and consumer outcome after the action.

Symptom guide

Symptom	Inspect first	Avoid assuming
One consumer is degraded	Consumer configuration, quota, and local dependency	The shared platform is healthy or unhealthy as a whole
Several consumers fail together	Shared service, resource pool, network path, or failure domain	Each workload has an independent fault
Intermittent latency	Saturation, contention, retries, dependency timing, or observation gaps	Average utilization proves adequate headroom
Change cannot complete	Capacity during transition, artifact identity, dependency readiness, or authorization	Retrying is harmless
Backup succeeds but restore fails	Recovery-unit selection, integrity, keys, dependencies, and verification	A successful job status proves recoverability
Management works but service fails	Consumer path and service-level outcome	Administrative reachability proves service health

Escalation packet

Provide the product and release identities, symptom and impact class, UTC time window, affected abstract failure domain, redacted health summary, recent approved change identity, and checks performed. Keep topology, addresses, tenant or customer names, credentials, raw configuration, and full logs in the approved private support system.

Use the cross-product troubleshooting method for ownership and closure.