Severity model
- SEV-3 — cosmetic issues, minor documentation drift or isolated stale labels.
- SEV-2 — repeated degraded health states, ingress/TLS metadata issues or registry update lag.
- SEV-1 — overview unavailable, broad ingress failure, or access behavior inconsistent with policy.
Routing guidance
| Issue pattern | Primary owner | Notes |
|---|---|---|
| Node pressure, scheduling drift | Platform / cluster operations | Check node details and related events first. |
| Ingress, routing or TLS symptoms | Networking / edge | Use service and observability views together. |
| Policy denial or access mismatch | Security / access owner | Compare expected path, role intent and endpoint behavior. |
Operator handoff
When escalating, include object name, namespace, affected image or node when relevant, the most recent warning pattern and the next path already checked. That shortens time-to-ownership significantly.