Shorthop Watch
Shorthop raises ShorthopIngestSilent when a Prometheus goes quiet — but that runs inside
Shorthop, so it cannot fire when Shorthop itself, its database or the whole cluster is down.
Shorthop Watch is a dead man's switch that runs on our side, outside your installation. A check expects a ping on a schedule; when the pings stop, the relay pages your phones directly and emails your contact address. It needs the push relay. The first check is free, then €1.99 per check a month.
- A Prometheus — Integrations → the Prometheus → Watch from outside. Shorthop forwards its Watchdog, so one check proves Prometheus → Shorthop → the network all work.
- Shorthop itself — create a check (e.g. every minute, 5 minutes' grace) and set its ping URL
as
HEARTBEAT_URL. The leader pings it only while it can actually send pages. - Anything else — backups, cron jobs:
your-job && curl -fsS -m 10 --retry 3 <ping URL>.
A ping URL is a credential: anyone holding it can keep a dead check looking alive. It is shown once; New URL replaces it.
What Watch sends outside your installation: the check's name, its state and two times, in the push and the email. Name checks with that in mind.