Uptime

Uptime monitoring

Watch a website or API from the outside — is it up, how fast, and when does its certificate expire.

Everything else in DevOps Agent looks at your server from the inside. Uptime looks from the outside, the way a visitor does: it asks your URL for a page, every few minutes, from somewhere else on the internet, and records what came back.

That is the difference between "the server is running" and "the website works" — and it is usually the second one you care about.

What it checks

  • Did it answer, and with the status code you expected.
  • How long it took.
  • Does the page contain the text it should (optional).
  • When the SSL certificate expires — read automatically on every HTTPS check.

Pages here

How a monitor decides something is down

By default a single failed check does not count: the monitor retries 3 more times and turns DOWN only if all fail (4 failed checks in a row). Networks hiccup; one timeout is not an outage. Set Retries before Down (0–10) per monitor.

Limits

50 monitors per project. The Free plan has 1 monitor in all; see Settings → Subscription for yours.

Alerts

When a monitor turns Down, an incident opens and you get an alert by email, Slack or Microsoft Teams, depending on the monitor's notification channels:

  1. 🔴 DOWN — straight away: the URL, when it went down and what the check saw.
  2. 🔎 Root cause — a few minutes later, if a server is linked to the monitor. Auto debug checks the server's disk, memory and SSH first, then the assistant looks at it in Read mode (it can look, not change anything) and sends a short summary of the likely cause. The full investigation is a normal chat you can open from the alert.

Then it goes quiet. While the incident is open you get no more alerts for that monitor, even if the site flaps or goes down again. Once the site is back up, click Mark fixed on the incident (with an optional note of what you did); the next outage opens a new incident and alerts again.

  • Where it goes: the monitor's Alert notifications, in its Edit form. It uses your team's default channels (at first, an email to the workspace owner) until you choose channels for it. Set the channels up in Settings → Notifications; the Notifications button on this page opens it. See Notifications.
  • SSL reminders: 14 days and 3 days before an https site's certificate expires, to the same channels.
  • Link a server in the monitor form (Auto debug → Server) to get the root-cause alert. Without one you only get the DOWN alert.
  • No alerts for one monitor: choose channels and untick them all. The list then shows No alert configured in red; the incident is still recorded.
  • Incident report: the header button exports every incident for the last 30 days to 12 months as PDF or CSV, with downtime, time to fix, the root cause and an SLA per monitor.

Not yet: there is no alert when the site comes back up, and no SMS or generic webhooks. Alert times are in UTC.

Who can do what

Everyone can see monitors, their checks and their response times. Operators and above can pause, resume, check now and mark an incident fixed. Only Owners and Admins can create, edit or delete a monitor, choose where its alerts go, and manage status pages. See Roles.