Uptime checks

Outbound probes — we hit your URL, port, or TLS endpoint on a schedule and alert when consensus across regions says it's down.

Three probe types

HTTP

We send a request, time it, and assert on status code, latency, and (optionally) a content match. Configurable method, headers, and body.

TCP

Connect-only check on a host:port. Great for non-HTTP services like Postgres, Redis, or your custom protocol.

SSL

Once-daily TLS handshake against a hostname. We alert when the cert is within 30, 14, or 7 days of expiry — three chances for someone to notice before it actually breaks.

Multi-region consensus

Each enabled region runs the probe independently on its own cadence. When you set a consensus threshold of N, a monitor only flips to failedafter N regions agree. That kills the “one flaky region wakes you up at 3am” failure mode without slowing down legitimate alerts.

The dashboard always shows the latest result per region so you can see where something started misbehaving even before consensus is reached.

Setting it up

  1. New monitor on the dashboard, pick the type (HTTP / TCP / SSL).
  2. Tell us the target — URL for HTTP, host:port for TCP, hostname for SSL.
  3. Choose the regions you want and the consensus threshold (we default to 2-of-N).
  4. Hook up alert channels and (optionally) an escalation policy.

Choosing the cadence

The default is 60 seconds for HTTP and TCP, daily for SSL. Faster cadences cost nothing extra on flat-rate plans but use API requests on free; we soft-cap at 30 seconds so probes can't DDoS your origin.