| docs

Appearance

Watchdog Dashboard ​

The Watchdog Dashboard provides real-time monitoring and health metrics for all services registered with the Watchdog service. It gives administrators a quick overview of overall system health and per-service availability.

Overview ​

URL: /admin/settings/watchdog-dashboard

The page lives under Settings > Tools > Watchdog and appears once the watchdog module is installed from Modules. Its services and settings tabs are documented on Watchdog Services and Watchdog Settings.

The dashboard summarizes the state of all monitored services and is the primary entry point for service health observability in AdminBolt.

Actions ​

  • Refresh - Reloads metrics and service status data.

Health Summary Cards ​

The top of the page displays a row of summary cards:

  • Overall Health - Aggregated system status (e.g., Healthy, Degraded, Unhealthy).
  • Monitored Services - Total number of services registered with Watchdog.
  • Total Checks - Cumulative number of health checks performed.
  • Average Uptime - Mean uptime percentage across all services.
  • Needs Attention - Number of services currently in a non-healthy state.
  • Auto Restart - Counts of services with auto-restart enabled vs disabled.

Settings Summary ​

A read-only summary of the active Watchdog configuration:

  • Watchdog Enabled - Whether the Watchdog service is active.
  • Check Interval - Frequency of service checks (in seconds).
  • History Limit - Maximum number of historical records retained.
  • Uptime History Limit - Maximum number of uptime history records retained.

To change these values, use Watchdog Settings.

Monitored Services Table ​

Below the summary, a table lists every registered service with the following columns:

  • Service Name - Name of the service (e.g., httpd, mariadb, postfix, sshd).
  • Status - Current state (Active, Inactive).
  • Uptime - Percentage uptime over the tracked window.
  • Checks - Total checks performed, broken down into active and inactive counts.
  • Auto Restart - Whether auto-restart is enabled for this service.
  • Last Checked - Time of the most recent check and the absolute timestamp.

Stuck web server workers ​

Since panel 1.9.0, on a server with Watchdog installed, a web server worker that stays stuck at full CPU for several minutes is detected and stopped automatically. Apache is not restarted, and the other workers keep running.

Each event is written to the panel log, as an error when the worker could not be stopped, and raises the System Notifications event for a stuck web server worker. The event is on by default and is delivered to the admin notification bell and through the notification channels you configured; see System Notifications.

From the server's command line:

  • bolt-watchdog stuck-workers history lists the detections. Add --limit to shorten the list or --json for machine-readable output.
  • bolt-watchdog stuck-workers settings get shows the detection settings.
  • bolt-watchdog stuck-workers settings set <key> <value> changes one of them, for example to switch detection off or to change how long a worker must stay at full CPU before it is stopped. Use settings get to see the keys.