Pulse — watchdog for Railway and Vercel
Polls your services, confirms an incident after N consecutive failures instead of alerting on the first blip, and on a Railway failure pulls the deploy log tail and names a likely cause right away.
- Incident confirmed after N consecutive failures, deduped with an occasional reminder, mutable for a set time
- A separate UNKNOWN status — a stale token doesn't look like an outage and won't close an open incident
- State lives in SQLite — a container restart doesn't re-blast alerts for known failures
- On a Railway failure — the deploy log tail plus a cause by schema: cause → evidence line → first action → confidence
- Control and alerts run through Telegram; restarting a service asks for a separate confirmation
Liked Pulse? We'll set it up for your business in 72 hours
No developers needed, implementation guaranteed
Deploys as its own service (Railway or any Python host) — not embedded in this site.