Health Monitor Guide
Quick operator reference for the Hub Health Monitor.
Status Meaning
healthyCritical checks passing
degradedOperational with non-critical issues
majorSignificant impact; investigate now
criticalCore runtime path broken
Core Checks
Worker/APIworker_health, worker_ops_config, surface_map, agent_health
Runtime/Controlruntime_status, runtime_show_state, control_show_state, runtime_sync
Topology/Framestopology_parity, frame_bytes
Cloud UIpages_routes
Blender Checks
blender_surface/blender-status reachability
blender_pipelineDetects frame contract mismatch, topology drift, stall, and colormap errors
Quick Triage
1Check overall_status, global_score, last_checked_at_utc
2Open incidents and inspect primary_cause + evidence
3Run /probes/run_once to force a fresh cycle
4Confirm 3 clean cycles to auto-resolve stale incidents
Full internal reference: docs/HEALTH_MONITOR_OPERATOR_GUIDE.md