Watch
All nine check types run on their own schedule, each with its own failure threshold.
Calmping watches your sites, APIs, workers, and databases from the outside — confirms a failure is real, opens an incident with the evidence attached, routes the alert, and hands you the timeline to fix it and the memory to prevent it.
Front end to database port, each check runs on its own interval with a failure threshold you set, so a single blip never pages you.
| Check | What it verifies | Example |
|---|---|---|
| Uptime | An HTTP request returns 2xx | GET /health → 200 |
| Keyword | The response body contains expected text | body ∋ "operational" |
| JSON field | A field in a JSON response matches a rule | status.ok = true |
| Latency | Average response time stays under a limit | < 300 ms / 10 probes |
| SSL / TLS expiry | The certificate has enough days left | alert under 14 days |
| DNS | A hostname resolves, optionally to an IP | example.com → A record |
| Heartbeat | A job or worker checks in on schedule | nightly backup pings |
| Database & TCPdatabases | A port accepts connections in time | Postgres · MySQL · Redis · Mongo |
| Custom HTTP | Any request plus your own assertions | signed, authenticated calls |
Authenticated checks bind an encrypted, app-scoped secret — Calmping can probe a private endpoint without that credential ever reaching another app.
Every outage travels the same path through Calmping. Nothing to wire together — the pipeline is the product.
All nine check types run on their own schedule, each with its own failure threshold.
A check crosses its threshold. Calmping confirms the failure is real before it makes a sound.
Email, Slack, Discord, or a signed webhook — queued, retried, and delivered. Browser push is coming soon.
The failure becomes an incident with its evidence attached. Recovery closes it on its own.
Each resolved incident drafts a five-field postmortem while the details are still fresh.
Confirmed postmortems become a per-app library that surfaces the fix that worked last time.
Email, Slack, Discord, and signed webhooks are queued in a durable outbox, retried after provider failures, and isolated — so one broken channel never blocks the rest. Browser push is coming soon.
Connect OpenAI, Anthropic, Gemini, or any OpenAI-compatible endpoint. On an incident, request a diagnosis — Calmping sends the incident class, evidence, and timeline, and stores the answer with the incident. Your keys stay encrypted, and nothing leaves without you asking.
Publish live service state, 30-day uptime, and incident history on an opt-in page — without exposing a single target URL, port, or secret. Your customers see health; your infrastructure stays private.
Monitoring means handing a tool your endpoints, tokens, and database ports. Calmping treats every one of them as a secret.
Every integration key and stored secret is encrypted with a master key that never lives in the database it protects.
RFC 6238 TOTP, with the enrollment QR rendered server-side — the secret never leaves Calmping for an external QR service.
An authenticated check's credential is bound to a single app. A slip in one place can't reach another.
Consistent database snapshots on the schedule you set, with configurable retention — no cron of your own required.
Connect an external heartbeat monitor to notice if the watcher itself ever goes quiet.
Export your configuration and run the whole platform privately in self-hosted mode when you'd rather own it.
Your first production service, fully monitored.
Faster eyes and room for a growing portfolio.
On the roadmap: additional monitoring regions for geographic failure confirmation. Monitoring currently runs from Frankfurt.