Monitoring services

Each service on your status page is one endpoint we check on a schedule. You manage them on the Services tab of your dashboard.

How we check

HTTP. We send a GET request to your URL.

• A 2xx response is operational.
• Any other status code, a connection error or no answer within the timeout counts as down.

The response body isn't read, so a simple endpoint that returns 200 OK is all you need.

gRPC. We call the standard gRPC health check (grpc.health.v1.Health/Check) on the host and port you give us, like grpc.acme.com:443. Use an https:// address for TLS.

• SERVING is operational.
• NOT_SERVING, a connection error or a timeout is down.
• Any other answer, such as UNKNOWN, is degraded.

Choosing an endpoint

Pick an endpoint that fails when your customers would notice: a /health route that checks your database is better than your homepage. It must be:

• Public. We check from the internet, so private IPs, localhost and internal hostnames are refused.
• Fast. Answers slower than your timeout count as failures.
• Free of secrets. Don't put tokens or passwords in the URL. URLs are never shown on your status page, but they're stored with your account.

Settings

From failed checks to incidents

A single failed check doesn't put an incident on your page. Short blips happen, and showing each one would make your page noisy.

• When checks keep failing for five minutes or more, that stretch becomes an incident. It ends at the first successful check.
• Each day's bar takes the color of its worst incident: red for down, amber for degraded, green when there were none.
• Uptime is the share of time without an incident.

Editing and removing services

Open a service from the list to see its recent checks and response times, or to change its name, URL or settings. Use Delete in the list to remove it. Its history is removed along with it.

Services marked Managed were set up for you by the Signalpost team. To change one of those, contact support.