Loading...
Loading...
A live read of the production health probe — updated every 30 seconds. Every value here comes straight from /api/health; we don't publish synthetic uptime numbers.
Scheduled probes sample the same /api/healthendpoint roughly hourly and store the result. Bars show recorded samples only — days before recording began say “no data”, they are never painted green.
No uptime history has been recorded yet. Sampling starts with the first scheduled probe and the bars fill in from that point — we don't backfill or synthesize uptime we didn't measure.
The live probe above and the hourly history sampler both run on the same deployment that serves this site — so a total platform outage can't record itself here. To cover that gap, an independent GitHub Actions probe on GitHub-hosted infrastructure also checks /api/health every hour and commits each result to the repository (status-history/probes.jsonl), outside our hosting provider entirely.
This page shows only what we can prove — the live probe and recorded samples. Nothing here is a synthetic uptime figure.
The uptime bars above come from stored probe results (raw JSON at /api/status/history). Incident history will appear here once there is a real incident to report — we don't publish an empty “no incidents” badge for a window we weren't recording.