USER MANUAL · SYSTEM

Health Check

Run 5-step diagnostics, review trust status, and export lightweight health evidence for your OpenClaw gateways.

/health-check

Overview#

Health Check provides a 5-step diagnostic wizard that validates every layer of your OpenClaw gateway — from basic connectivity to IM channel readiness. It runs as part of the onboarding flow when you first add a connector, and remains available as a standalone tool for ongoing monitoring.

Beyond the diagnostic wizard, Health Check keeps a lightweight evidence trail: trust status from nightly TCC audits, the recent health-check timeline, and exportable summaries for incident review. Together, these signals give you confidence that your Agent infrastructure is operating correctly without introducing a separate AI report product surface.

Prerequisites#

Prerequisites

At least one connector configured (see Connectors)

Expected Results#

After running a health check, you will see:

-5-step diagnostic results — each step shows pass (green), warning (amber), or fail (red) with specific remediation suggestions
-Overall status summary — healthy, degraded, or critical based on the combined step results
-Trust score — a TCC (Trusted Change Control) nightly audit summary reflecting configuration safety
-Health check history timeline — past results for trend tracking and regression detection
-Lightweight health evidence — trust snapshots and recent diagnostic summaries you can export or review later

Web Operations#

Navigate to Health Check from the System section in the sidebar.

Running a Health Check

Click the Run button to start a new diagnostic. The wizard executes 5 steps in sequence: (1) Connectivity — verifies the gateway is reachable and checks for secure connection (wss/https), (2) Model Availability — discovers agents and validates their health status, (3) Cron Sandbox — checks scheduled job success rates, (4) Budget Protection — verifies cost protection configuration, (5) IM Channels — probes Telegram, Discord, and Slack for config, login, and messaging capabilities. Each step completes independently, so you get partial results even if later steps fail.

Understanding Results

Each step displays a pass, warning, or fail status. Pass means the check completed successfully with no issues. Warning indicates a non-critical issue that deserves attention — for example, a cron job with a low success rate or an IM channel that is configured but not logged in. Fail means a critical problem that needs immediate action, such as the gateway being unreachable. Each warning and failure includes a remediation suggestion explaining what to do next.

Trust Status

The Trust Status panel shows the TCC (Trusted Change Control) nightly audit summary. This is a background process that runs daily, checking configuration drift, permission changes, and security posture. The panel displays the latest trust score and any flagged issues from the most recent audit cycle.

Health Check History

All past health check results are stored and displayed in a timeline view. Each entry shows the date, overall status, and per-step results. Use the history to track trends — for example, spotting a connectivity issue that recurs at specific times, or confirming that a remediation action resolved a previous warning.

CLI Operations#

Show the current health status of your connected gateways

Terminal

$ ap health-check status

Run a new 5-step diagnostic check against your gateways

Terminal

$ ap health-check run

Run a full diagnostic with detailed remediation suggestions for each issue

Terminal

$ ap health-check diagnose

Mobile Operations#

On mobile, go to the More tab and select Health Check. The mobile version provides a simplified readiness check view.

-Simplified 5-step diagnostic view with pass/warning/fail indicators per step
-Overall gateway status summary (healthy, degraded, or critical)
-Pull-to-refresh to re-run the readiness check on demand

FAQ#

Q: Health check shows 'warning' but everything seems fine.

A: Warnings flag non-critical issues that may not affect daily operations but could cause problems later. Common examples: a cron job with a below-threshold success rate, an IM channel configured but not logged in, or a missing budget policy. Review the remediation suggestion for each warning — you may decide the warning is acceptable for your setup, or you may want to address it proactively.

Q: How often should I run health checks?

A: There is no strict schedule. We recommend running a check after any infrastructure change (gateway update, network migration, new agent deployment) and periodically (weekly or monthly) for routine verification. The Trust Status audit runs nightly automatically, so you always have a baseline, and the recent history helps before important deployments or incident review.

Q: What replaced Doctor Digest?

A: The simplified control plane keeps the diagnostic wizard, trust-status audit, and lightweight history/export evidence, but it no longer promotes a separate AI-generated digest product surface. If you need a deeper write-up, export the evidence and review it in your own incident workflow.