API monitoring that catches failures before your users do
An endpoint can return 200 and still be broken. Check every API, public or authenticated, every minute, validate the response, and get alerted the moment something drifts.
Monitor private, authenticated APIs
Add custom request headers to any check (Authorization bearer tokens, API keys, whatever your endpoint expects) so internal and private APIs get watched too.
GET, HEAD, and POST checks
Pick the method that matches the endpoint, set a timeout, and choose whether redirects should be followed. Health checks, webhooks, and form endpoints all fit.
Validate the response, not just the connection
Define the status-code range you expect and check the response body for a keyword, present or absent. A 200 that returns an error page still counts as down.
Degraded, not just down
Set a response-time threshold and slow-but-up endpoints show as degraded instead of down, so you can spot trouble before it becomes an outage.
How it works
- 1
Add your endpoint
Create an HTTP monitor with the URL, method, and any request headers your API needs, auth headers included.
- 2
Define what healthy means
Set the expected status-code range, an optional keyword check on the body, and a degraded response-time threshold.
- 3
Route your alerts
Choose SMS, email, or Discord per monitor: down alerts, recovery notices, or both, with customizable message templates.
Frequently asked questions
Can I monitor authenticated or private APIs?
Yes. Every HTTP monitor supports custom request headers, so you can send an Authorization bearer token, an API key header, or anything else your endpoint requires to respond.
Which HTTP methods do checks support?
GET, HEAD, and POST. Each monitor also has its own timeout and a follow-redirects toggle, so you can match the check to how the endpoint actually behaves.
How do I verify the API returns the right response?
Set an expected status-code range and add a keyword check on the response body: require a keyword to be present, or flag the check as failed if it appears. An endpoint that responds with the wrong payload is treated as down.
What happens when an endpoint is slow but still responding?
Set a degraded response-time threshold on the monitor. Responses slower than the threshold mark the check degraded instead of down, so you see performance trouble without false outage alerts.