There are three states a service check can be in: up, down, and slow.
How do we define down
A check will be considered down if we are unable to connect to it, resolve the URL/IP address or if there is some other error.
Errors could be nameserver lookup failures, timeouts, or SSL certificate errors.
A retry is attempted when a check is found to be down to verify the state. Timeouts are capped at 20s, regardless of what is entered in your portal.
A down status will not be recorded for error pages or HTTP status codes, e.g. a 500 internal server error will be considered "up" because the check is responding. To be notified in this situation, you can create a Content not Found or HTTP status code alert check.
The "down" status affects the uptime % calculations.
How do we define up
A check will be considered up if we receive any status from a check. A 5xx error message still will be considered up. You can configure an alert based on 5xx status codes if this is something you would like to be notified about.
How do we define slow
Slow is an extra state which helps with reporting and alerting when services are responding slower than you would expect.
A check is considered slow if any response time greater than 500ms.
This can be configured from the "Fine Tuning" section when adding or editing a check.