| Alerting pipeline down | The always-firing Watchdog alert is missing. Prometheus is not evaluating the rules or cannot reach Alertmanager, so alerts that should fire may not be listed. If nothing else is listed, the page says that this is not all clear. | Watchdog runbook |
| Alert source unreachable | No Alertmanager member answered the control plane. No list is shown. The message names each member and its error. | Check docker compose ps in /opt/veristar/observability on the core nodes (alerts runbook) |
| N of M Alertmanager members did not answer | Some members are down. The list comes from the others; every member holds every alert, so it is complete. | Bring the named member back |
| No alert source configured | The control plane has no ALERTMANAGER_URLS. Deployments made with the Ansible roles always set it. | Contract section 11.3 |