Public availability · external probe

uptime over the last 24 hours

/health is probed every minute from outside the bot's own stack — so a crashed service produces a real failure-recording, not a silently-correct response. This page is hydrated from /api/uptime, which rolls the probe log into windows. The minute-bar timeline below shows the last 24 hours individually.

24 hours
probes loading
7 days
probes loading
30 days
probes loading

Last 24 hours, one bar per minute

Loading…
OK (HTTP 200) Degraded (4xx/5xx) Error (timeout/refused) No probe data