The Service Watchdog monitors all mail server services and alerts you when something goes wrong. A persistent banner appears at the top of every dashboard page when any service is down, and email notifications are sent to the admin when mail-critical services stop running.
Email alerts for service outages are available on the Pro plan and above. The dashboard banner showing service status is available on all plans.
Dashboard banner
When any monitored service stops running or fails, a banner appears at the top of every dashboard page. The banner shows:
- Which service is down and its current status (stopped or failed)
- A Restart button to immediately restart the service
- Dismiss to hide the alert for 30 minutes
- Don't show again to permanently hide alerts for a specific service (useful for optional services you don't use)
The banner updates in real time and automatically disappears when the service recovers. Temporary dismissals reset when you log in again.
Email alerts
The watchdog sends an email to the primary admin when any monitored service goes down. Monitored services include:
| Service | Role |
|---|---|
| Postfix | Sends and receives email via SMTP |
| Dovecot | Provides mailbox access via IMAP |
| OpenDKIM | Signs outgoing email for DKIM authentication |
| MariaDB | Stores domains, users, aliases, and configuration |
| SpamAssassin | Filters inbound spam |
| SpamAssassin Milter | Milter bridge connecting SpamAssassin to Postfix |
| Nginx / Apache | Web server for dashboard and webmail |
| Unbound | Local DNS resolver |
| Rsyslog | System logging |
| Fail2ban | Brute-force protection |
| CeyMail AI Filter | AI-powered email screening (if enabled) |
How it works
The health watchdog runs every 30 seconds via a systemd timer. When it detects a service is down:
- Sends an email to the primary admin with the service name, status, and server hostname
- Marks the alert as sent to prevent duplicate emails for the same outage
- Clears the alert when the service recovers, so future outages trigger new notifications
Alert email
The alert email includes:
- Which services are down
- The server hostname
- The time of detection
- A prompt to log in and investigate
If your server has an SMTP relay configured (Amazon SES, SendGrid, etc.), the alert email is sent through the relay. Otherwise, it uses direct delivery. Even if Postfix is the crashed service, the email is queued locally and delivered once Postfix is restarted.
Dismissing alerts
Temporary dismiss
Click the X button on any service alert to hide it for 30 minutes. The alert reappears after 30 minutes or when you log in again. This is useful when you're already aware of an issue and working on it.
Permanent hide
Click the eye-off button to permanently hide alerts for a specific service. This is useful for optional services you intentionally keep stopped (for example, the AI Filter if you don't use AI screening). Permanent hides persist across logins and page refreshes.
To restore hidden service alerts, clear your browser's local storage for the dashboard domain.