mcp.kituwa.com · Watchdog
What is running, when it last checked in, and what failed. Six checks, computed for this request and cached for at most 60 seconds. No hardcoded green.
State
Current state only. No history store on this light profile. Current state is measured; a percentage would be invented.
Schedule
- Mode
- Request driven, no scheduler
- Checks
- Checks recompute on request and are cached for at most 60 seconds. The registry is refreshed by the estate orchestrator; this site never writes it.
- Previous
- computed for this request · 3s ago
- Next
- next request after the 60 s cache window
- Tracking since
- 2026-11-21
- Registry
- last written 17m ago by the estate orchestrator (this site never writes it)
- Deploy
- app/main.py last modified 9m ago; version 1.0.0
Weekly progress
Each tile is one local day. Grey means no run was recorded, because this light profile has no scheduler and no history store. Nothing is inferred from a missing row.
Checks
Every row shows its own evidence and check time (2026-11-21 22:19:36 UTC). A row is only green when the check actually succeeded; "unknown" is used when a check cannot run in this environment rather than being counted as a pass.
Recent runs
No runs recorded. This site has no scheduler: checks exist only as request-time measurements, so there is no run log to show. The unit that would schedule them lives in deploy/systemd/ and is not installed on the light profile.
Uptime and failures
No history store on this light profile. Current state is measured; a percentage would be invented. Current failures: 0. Degraded or unknown: 0. Machine-readable: /watchdog.json.
What this covers
Covers: the app answering and the registry loading, the TLS certificate of the public origin, the reverse-tunnel systemd unit, the local mouse-driver health endpoint, registry file freshness and entry count, and deploy artifact freshness. Does not cover: a real operator sign-in by a person, the mouse driver's motion correctness, uptime percentages, or any other host.