status.arc42.org arc42 section 12
Glossary
The glossary table was still empty placeholders in the original documentation. The terms below are the ones actually used and defined across the docs and ADRs.
| Term | Definition |
|---|---|
| Vantage point | The network location a probe request originates from. status.arc42.org currently has a single vantage point (one GitHub-hosted runner region); a second could be added without a schema change. |
| Honesty chain | The three-layer design that prevents an outage from failing silently: a static error panel when the backend is unreachable, a stale status when probe data stops advancing, and a Slack alert when a check fails. |
| Probe run | One execution of the availability prober, covering all monitored sites. Recorded as a single probe_run heartbeat row regardless of outcome — the signal the freshness check reads. |
| Status bucket | A per-site, per-day rollup of downtime minutes and outage count, upserted on every probe run. What the 30-day history bars render from, so rendering never scans the raw event log. |
| Status snapshot | A recorded state transition for one site (e.g. up → down), written only when the state actually changes. |
| Confirmed outage | A site is only marked down after 2 of 3 probe attempts, five seconds apart, fail — preventing a single dropped packet from being recorded as an incident. |
| Plausible.io | Commercial, cookie-free web analytics service used to collect visitor and pageview counts, chosen for its EU hosting and GDPR/CCPA/PECR compliance. |
| Turso | A cloud-hosted, distributed SQLite database (libSQL), used to persist availability data, rate-limit timestamps, and startup metadata. |
| Atlas | A declarative database schema management tool; the desired Turso schema is described in HCL and kept in sync with the Go code manually. |
| cron-job.org | The external cron service that triggers availability probing every 15 minutes via POST /api/probe, chosen over GitHub Actions cron for independence from the systems it monitors. |