Service
Website uptime monitoring
We request each monitored URL (by default every 10 seconds) and record the HTTP status code and response time.
What the statuses mean
- Healthy
- The URL returned an accepted status code (by default 200–399). You can set your own accepted codes, such as
200or200,301,401. - Degraded
- The server answered, but not as expected: another status code (for example 503, or 401/403 reported as “access restricted”), missing text, a slow response or a redirect that could not be followed.
- Down
- No usable HTTP response: DNS failure, connection refused, timeout or TLS handshake failure.
- Unknown
- No result for three check intervals. Old data is never shown as healthy, and missing data is never shown as downtime.
- Paused / Maintenance
- Paused by you or because monitoring is inactive (with the reason), or inside a maintenance window.
Certificates do not decide availability
Reachability (did any HTTP response arrive?) is shown next to the state. The probe reads the HTTP response even if the certificate is expired, self-signed or issued for another name. That problem is reported separately under SSL & Domains, so a working site with a certificate problem shows Healthy plus a certificate Invalid state.
Checks run from one location (our monitoring server). Outages shorter than the check interval, or visible only from other networks, may not be detected.
Redirects
By default a redirect (3xx) response is judged by the accepted status codes and the destination is shown but not requested. You can enable redirect following per website; each destination is validated before it is contacted, up to five hops.
Incidents and alerts
You are alerted as soon as a website goes down. If it stays down, reminders follow 5 minutes, 15 minutes, 5 hours and 24 hours after the outage began, each showing how long it has been down. Reminders stop the moment the site recovers, and you receive one recovery alert with the total downtime. You can require several consecutive failed checks before an outage is declared, to filter out one-off network blips. Alerts are never duplicated.
Fair use and safety
Only public websites on standard ports (80 and 443) can be monitored. You verify ownership of each domain with a DNS TXT record or a small file. Private, internal and cloud-metadata addresses are always blocked.