Uptime, response time and SSL monitoring
Website Monitor requests your URLs on a schedule, records what came back, and tells you when the pattern means something is really wrong.
What a check does
Each monitor is an HTTP or HTTPS request that you configure once.
- HTTP and HTTPS
- Monitor any public http:// or https:// URL, using a GET or HEAD request.
- Scheduled checks
- Choose 1, 5, 10, 15, 30 or 60 minutes. The shortest interval available depends on your plan. A new or resumed monitor is checked right away, and you can run Check now at any time.
- Status codes
- Decide which response codes count as healthy: a single code, the 2xx range, or a custom range.
- Timeouts
- Set how long to wait for a response, from 1 to 60 seconds. A request that times out is a failed check.
- Response time
- Every check records how long the response took. A successful response slower than 3 seconds is shown as degraded rather than down.
- Pause and resume
- Pause a monitor during planned work. Time spent paused is not counted as downtime.
Failure and recovery thresholds
The rule that keeps alerts meaningful.
Every monitor counts consecutive results. After 3 failed checks in a row (the default; you can choose 1 to 10) an incident is opened and your down alert is sent. A single failure that is followed by a pass changes nothing.
Recovery is evaluated separately. After 2 passing checks in a row (the default; also 1 to 10) the incident resolves, the total downtime is recorded and your recovery alert is sent.
Incidents move through Investigating, Identified, Monitoring and Resolved. The automatic messages state what the monitor observed, for example the HTTP status it received or that the host was unreachable, and never a guessed cause. You can publish your own updates along the way.
SSL certificate monitoring
Tracked on every HTTPS check, independent of uptime.
- Expiry awareness
- The certificate's issuer and validity dates are read on each check, and the days remaining are recorded.
- Expiry thresholds
- You are alerted when the certificate has 30, 14, 7 and 3 days left, once per threshold, and again if it expires. A renewed certificate starts a fresh set of alerts.
- Separate from uptime
- A site can be fully operational with a certificate about to lapse. That is reported as a warning, not an outage.
- Certificate problems
- If a TLS handshake fails, the check fails with a clear certificate error that counts toward an incident like any other failure.
Notifications
Tell the right place, once.
- Email and webhooks
- Add email recipients and webhook endpoints under Settings. Webhook requests can be signed with a secret that is stored encrypted.
- Choose the events
- Each channel can be switched on or off for down alerts, recovery alerts and expiry alerts (SSL certificate and domain registration).
- No duplicates
- An event is delivered once per channel. Each alert is recorded before it is sent, and temporary failures are retried with backoff.
- Send a test
- Send a test email from Settings to confirm that delivery works before you rely on it.
Alerts travel through third-party email and network infrastructure, so delivery cannot be guaranteed. See the Terms of Service.
Current limits
What is not covered today, so you can judge the fit.
- Checks are HTTP and HTTPS requests. There are no ping, port or keyword checks.
- Checks run from one monitoring location, so a result reflects the view from there.
- Only public addresses can be monitored. Private, internal and cloud-metadata addresses are refused.
- Alerts go to email and webhooks. There are no SMS or chat-app integrations yet.
Know when something goes wrong.
Start monitoring your websites before your users have to tell you something is broken.