knell is a two-way watch. It rings a bell (a push notification) the moment something it is watching goes quiet or unreachable — a self-hosted blend of a dead-man's-switch and an uptime probe. The counters above are every check it is currently holding.
Your job pings knell on a schedule. Miss the window and knell tolls. Best for cron jobs, backups, pipelines — anything that should phone home.
knell fetches your endpoint on a schedule. Enough failures in a row and knell tolls. Best for sites and APIs that should answer.
Operators register checks and watch the live board on the dashboard (access-controlled). Each check gets a private link you can bookmark or share.