Monitoring HTTP endpoints

An HTTP monitor checks that a URL returns the response you expect, on a schedule, from the regions you choose. This guide covers how the check works and how to configure it.

How it works

Stector sends a request to your monitor's URL on the interval you set, from each region you've enabled. Each individual run either matches your expected status code or it doesn't — there's no retry of the request itself. Except on a new or just-resumed monitor's first check, a single failed run doesn't mark a monitor down; the rule is under Regions and intervals below.

Configuring the request

From /dashboard/monitors/new, choose HTTP monitor and set:

  • Method — GET, POST, PUT, DELETE, or HEAD.
  • URL — the endpoint to check.
  • Expected status — the exact status code a successful run must return, for example 200. Stector compares this as an exact match, not a range: a monitor expecting 200 treats a 204 response as a failure.

Once a monitor is created, its type can't be changed. On the edit page at /dashboard/monitors/[id]/edit, Monitor type appears as a read-only label instead of a dropdown.

Authentication

Under Authentication, choose one of four types: None, Bearer token, HTTP Basic, or Custom header. Basic adds a Username field alongside the secret; Custom header adds a Header name field (for example X-API-Key) alongside the secret.

If a monitor has authentication configured and the endpoint responds with a redirect, Stector treats that as a failure rather than following it — forwarding an Authorization or custom header to a redirect target could send your credentials somewhere you didn't intend. Point the monitor's URL at the final destination instead.

When editing a monitor that already has a secret saved, leave the secret field blank to keep the existing value. Stector only replaces it when you type a new one.

Regions and intervals

Monitors run from three regions — US East, EU West, and Asia — on a Check interval from 1 minute (3 minutes on Free) to 1 hour. Once a monitor is running from more than one region, a failure must be seen twice in a row from a region, and two regions must agree, before it pages you. That has limits:

  • Stector is set up not to page you for an outage visible from only one of a monitor's regions. The exception is a new or just-resumed monitor's first check (see the last item).
  • The three regions are in Ashburn, London and Singapore. There's no coverage from Oceania or South America.
  • Any monitor set to one region needs two failed checks in a row from that region. That includes every Free monitor and any paid monitor set to one region.
  • A new or just-resumed monitor's first check decides, in either direction, from one region.

Stector is set up to mark a down monitor back up once the latest run in enough of its regions is a success. Managing incidents gives the exact numbers.