How checks work

• Checks run on your monitor’s interval from Cloudflare’s global network.
• Retries before alarming: a single failure doesn’t immediately mark a monitor down. A failing check is retried up to 3 times (about 12 seconds apart) to rule out a transient blip. Only if it’s still failing does the monitor flip to down and open an incident. (A manual “check now” retries more quickly.)
• Response time is recorded on every check and charted over time.
• Incidents open automatically when a monitor goes down and resolve automatically when it recovers — no manual acknowledgement needed.
• Monitors in maintenance are still checked for live status but do not record incidents or send alerts.