Context
With a growing fleet of services, 'is it down?' needed a single honest answer. The status system combines external monitoring with self-hosted probes and a public page.
Build
Fastify gateway aggregates checks; Python probes run on the host; a durable Telegram relay makes sure alerts actually arrive.
The public page answers users; the retention and backups answer post-incident questions.
TODO(user): incident stories worth telling.
Facts
- Public status page plus internal gateway API
- Host probes in Python on cron; Telegram relay with durable SQLite retry queue
- 730-day incident history retention with daily backups
- Monitors every public Okiyumi service