The public endpoints of Chat Guard are checked every few minutes by an external monitor, with GitHub-based probes as a backup; an outage alerts the operator.
This repository is that backup. probe.sh runs from GitHub Actions on GitHub's schedule, which is
best effort, so runs can be hours apart. It checks:
https://api.chatguard.dev/readyz(API and its database and cache)https://api.chatguard.dev/openapi/v1.jsonhttps://app.chatguard.dev/sign-in(dashboard)https://chatguard.dev/(site) and its/terms,/privacyand/dpapageshttps://www.chatguard.dev/(must redirect tohttps://chatguard.dev/)https://chatguard.dev/discord(must redirect to our Discord invite)- the TLS certificates of the four hosts (must have at least 10 days left)
Each target gets three attempts twenty seconds apart. A new outage opens an issue labelled incident, and the failed run e-mails the repository owner; recovery closes the issue with the duration. Open and past incidents: issues labelled incident.