System Messages
Health check alerts and infrastructure consistency notifications
What you see
URL: /system/messages
The system messages page shows health check results detected by PodWarden's periodic infrastructure audit. PodWarden automatically checks for data inconsistencies between its database and your live infrastructure every 15 minutes.
A bell icon in the top navigation bar shows the number of unread messages. The badge color reflects the highest severity:
| Color | Meaning |
|---|---|
| Red (pulsing) | Critical issue — e.g. cluster unreachable |
| Red (static) | Error — e.g. node name mismatch preventing canvas visualization |
| Amber | Warning — e.g. host offline, storage degraded |
| Blue | Informational — e.g. stale assignment |
Click the bell to open the system messages page.
Health checks
PodWarden runs these checks automatically:
Node naming
| Check | Severity | What it means |
|---|---|---|
| Node name mismatch | Error | The K8s node name stored in PodWarden doesn't match the actual node name in your cluster. This prevents deployment-to-node connections from appearing on the canvas. |
| Node name missing | Warning | A host is assigned to a cluster but has no K8s node name recorded. |
Cluster membership
| Check | Severity | What it means |
|---|---|---|
| Cluster unreachable | Critical | PodWarden cannot connect to the Kubernetes API for this cluster. |
| Orphaned host | Warning | A host claims to be in a cluster, but the cluster doesn't have a matching node. |
| Unknown K8s node | Error | A node exists in your K8s cluster that PodWarden doesn't know about. |
Storage
| Check | Severity | What it means |
|---|---|---|
| Longhorn pods not ready | Warning | Some Longhorn pods are not running. Storage may be degraded. The message lists which pods are not ready. |
| Longhorn node not schedulable | Warning | A Longhorn storage node cannot accept new volume replicas. This usually means the disk is too full — Longhorn requires at least 15% of the disk to be free (configurable via storageMinimalAvailablePercentage). Free up disk space or add more storage to resolve. |
Deployments
| Check | Severity | What it means |
|---|---|---|
| Stale assignment | Info | A deployment references a cluster that no longer exists. |
| Unplaceable deployment | Warning | A deployment's placement targets a node that no host matches. |
Hosts
| Check | Severity | What it means |
|---|---|---|
| Host unreachable | Warning | A host hasn't reported stats in over 30 minutes. |
| Tailscale hostname drift | Info | A Tailscale-discovered host is missing its hostname. |
Using the page
Filtering
- Severity chips at the top toggle which severity levels are shown
- Category dropdown filters by check category (node naming, cluster, storage, deployments, hosts)
- Status toggle switches between active issues, resolved issues, or all
- Click a message row to expand and see full details and diagnostic data
Actions
- Mark as read — dismisses the notification badge for you (other users still see it as unread)
- Mark all read — marks all active messages as read
- Run check (admin only) — triggers an immediate health check instead of waiting for the next cycle
- Delete (admin only) — permanently removes a message
Auto-resolution
When a previously detected issue is no longer found during a health check (e.g., you fixed a node name mismatch or freed up disk space), the message is automatically marked as resolved. Resolved messages are cleaned up after 7 days.
Email notifications
PodWarden can send email alerts when new infrastructure issues are detected. Configure this in Settings → System Config → Notifications.
| Setting | Description |
|---|---|
| Enable email notifications | Master toggle |
| Recipients | Comma-separated list of email addresses |
| Minimum severity | Only issues at this severity or above trigger an email |
Emails are sent as a digest — one email per health check cycle containing all newly detected issues. Previously known issues do not re-trigger emails. If an issue resolves and later reappears, it is treated as new.
Email delivery uses the SMTP settings configured in the same Settings page. SMTP must be configured and working for email notifications to function.