API monitoring
API monitoringthat reads the answer, not only the status.
Know It Broke calls your API or website as often as every minute and judges what comes back. When it is wrong twice in a row, the right person hears. When it is right again, they hear that too.
The request
Call it the way your own code does
Pick the method, add your own headers and send your own body. That covers health endpoints, signed-in endpoints behind a key, and form posts.
Press "Test it" before you save and you see the real answer, the time it took and whether your checks pass.
- GET, POST, PUT, PATCH, DELETE or HEAD
- Your own headers and request body
- Secret values stored encrypted and shown as dots
- As often as every minute
- Test it before you save it
- The status code
- Text that must be in the body
- A value in the JSON
- A response header
- How long the answer took
The checks
A 200 is not the same as working
An API can answer 200 with an error inside, or an empty list where the orders should be. Add checks for what a good answer looks like.
Every run is kept for 30 days with its timing and the reason it failed, so you can see when it started going wrong.
The alert
Two bad answers, then somebody is told
One failed check could be a blip, so nobody is woken for it. A second failure a minute later means down, and the alert goes out.
Alert groups choose who hears by time of day: the team channel in office hours, whoever is on call at night.
- Slack, Discord or any webhook
- PagerDuty
- Text message and email
- A call to your own API
- A message when it is back, and how long it was down
Questions
How often can an API be checked?
As often as every minute. You choose the gap for each node, from one minute up to an hour.
Can it check an API that needs a key or a token?
Yes. Add the header your API expects. Secret values are stored encrypted and shown as dots once saved.
Where do the checks come from?
Cloudflare's network, from the place you pick for each node: the United Kingdom, Germany, US East, US West, Singapore or Australia. New nodes start in the UK. If your API sits behind a firewall, the Setup page lists the address ranges to allow and the name the checks send.
Does it check more than websites and APIs?
Yes. The same account checks ports, DNS records and scheduled jobs, and on the Standard and Plus plans runs browser flows.
Also in every account