Troubleshooting
Start with the status, the last-check time, and the check details. Then compare the result with the free tools from the same public perspective.
The endpoint is Unreachable
An error means the check itself could not complete, for example because of a refused connection, DNS failure, timeout, a mail or database server that does not offer STARTTLS, or a service CertWatch does not read, such as MySQL (see step 2 of the setup guide). A failed check appears on the endpoint's Monitors row and monitor card as "Host unreachable since …" beside its last status, and on the Dashboard's Triggered list when that status is not Healthy, but does not dispatch an alert, because transient network failures are common. CertWatch normally checks again about two minutes later, and when a second check fails too the endpoint becomes Unreachable and sends an Unreachable alert, then again every 24 hours while the checks fail. Investigate the target, then confirm recovery with a fresh check.
A check that reads "Not conclusive · CertWatch could not finish the check" failed on CertWatch's side: its DNS lookup got no answer, its time ran out before it could try any address, CertWatch had no IPv6 route to an IPv6-only endpoint, or fetching the certificate's issuer timed out. It never counts against your endpoint and sends no alert of its own: the endpoint keeps the status its last reading gave, moved up as that certificate's Late, Urgent and expiry dates pass, which still alert. An endpoint already Unreachable after an Unreachable alert stays Unreachable, and one CertWatch has never read stays Unreachable with that reason until a check finishes.
The certificate is Late, Urgent, or Down
These are the Late, Urgent, and Down alert types. A Recovered alert follows when a later check finds the certificate Healthy, if the problem sent an alert, unless the endpoint is muted then. The monitor card's Overview tab lists what the last check found.
- Late or Urgent because of the renewal: the renewal did not happen when this endpoint's renewal was expected. Check the renewal client's log to see whether it ran and succeeded, compare the serial served with the certificate on disk (a renewal that was never deployed looks the same from outside), confirm the server reloaded after the renewal, and check the CA's status page for an issuance or validation outage. See Certbot renewal failed and Old certificate after renewal.
- The renewal dates are wrong for this endpoint: if the certificate is renewed by hand, or by automation with a set number of days left, set that on the monitor card's Settings tab (see Renewal mode). On an active endpoint CertWatch checks it again with the new setting, so its dates and status follow within seconds; a paused endpoint has none until you resume it. A new status that check finds sends its alert at once.
- Urgent or Down because of the certificate itself (a missing intermediate, a weak key, a name it does not cover): the alert names the finding and what to do about it. Reissue or reinstall the certificate with the complete chain, then verify the deployed result from outside with the checker. See Incomplete certificate chain.
- You already know the renewal is coming: use Acknowledge to hold the reminders.
A new certificate sent no alert
A new certificate, issuer, key, or set of names is what a healthy renewal looks like, so CertWatch records it as a "New certificate observed" entry on the monitor's History and sends nothing. If the change was not yours, confirm whether a deployment, CA migration, or proxy change explains it. A new certificate that breaks something, such as a missing intermediate or a name it no longer covers, alerts by its status.
The domain is unsupported or has unexpected registration data
For some TLDs CertWatch has no RDAP source that publishes an expiry date. CertWatch also never reads internationalized (IDN) TLDs. That is a coverage gap, not proof that the domain is down. Supported TLDs lists which TLDs are covered and why. For supported domains, compare the registration result with the registrar's authoritative account and confirm the renewal date and transfer settings there.
No alert arrived
- A certificate that renews on schedule never alerts. Only a late renewal, a certificate problem, or failing checks do.
- A single failed check shows only in the app. An endpoint whose follow-up check about two minutes later fails too sends an Unreachable alert, and a domain after 24 hours.
- Confirm the alert email is verified and current.
- Check whether the endpoint is muted. A mute holds every alert, the Recovered included, but does not stop checks or history.
- Check whether the endpoint's alert is acknowledged, in the Alert state section of its monitor card's Overview tab. No alert listed there, or no section at all, means it has no open alert.
- Check the Alert channels card on Alerting: a severity can be routed away from a channel, or to none.
- For Slack, confirm the plan supports it and that the Slack tab's Connection section shows a channel. If alerts fail with a rejected connection, use Reconnect Slack; if the channel is gone, choose another. A private channel needs
/invite @CertWatch. - For generic webhooks, confirm the plan supports the channel and that the destination still accepts requests.
- Review the endpoint's last check in the app and alert history before treating the absence of an email as proof that monitoring stopped.
Email support@certwatch.sh, or use Support in the app.