# FAKAP — service watchdog for the virya stack

> FakApp is a small, single-binary watchdog that probes public health
> endpoints of the virya production services and emails the crew when one
> stops answering. This page is its read-only status board.

## What this board shows

- One card per monitored target: current state (up / down / unknown),
  latest successful latency, uptime share over the retained probe history,
  how long the target has been in its current state, and a strip of recent
  probe outcomes.
- A banner that reads ALL SYSTEMS GO only when every target is up.

## How detection works

- Each target is probed with HTTP GET on a fixed interval (default 30 s).
- State changes need consecutive failures (default 3) or consecutive
  successes (default 2), so a single blip is not an outage.
- Every down-transition sends an email; while a target stays down, a
  reminder is re-sent after a fixed quiet period. Recovery sends exactly one
  all-clear.

## Machine-readable metadata

- `/llms.txt` — this file.
- `/healthz` — plain-text liveness of the watchdog process itself.
- `/style.css` — the board stylesheet.

## Scope and authority

FakApp observes; it never mutates anything it watches. It holds no
credentials for the services it monitors and has no write path into them.
Business truth stays in CrowdRelay; this tool answers exactly one question:
"is it reachable right now?"