Skip to content

Where do I see a monitor's incidents?

An incident is a time a monitor was Down: from its first failure, once two of our checking places agreed on it (or one place failed three times in a row when the other could not check), until the monitor answered as it should again, or until it ended another way. Each incident keeps its cause, its times, the places that saw it and the emails it sent, for as long as your plan keeps history.

On a monitor's page

Under the numbers at the top, Incidents shows the monitor's 5 latest: their Status, their Cause, when they Started, and how long they lasted. Click a cause to see that incident in a panel at the side, without leaving the monitor's page. All incidents opens the Incidents page with all of this monitor's incidents your plan shows.

The Incidents page

Open Incidents in the side menu. The page opens on the Last 7 days.

  • Ongoing comes first: every incident still going on, whatever time range you chose.
  • Ended lists the incidents that are over and were down during the time range, also one that started before it, newest first, 25 a page. Click Started or Duration at the top of a column to sort by it; click again to turn the order around.

Each row shows:

  • Status: Ongoing while the monitor is still down, Resolved once it is up again. An incident that ended another way says how: Can no longer check, Ended: paused, Ended: check switched off or Ended: check changed.
  • Cause: what our checks found, in the words of the Down email, such as Server error (502) or Port closed.
  • Started, Ended and Duration, in your time zone.

Click a row, its cause or View at its end to open the incident in a panel at the side, without leaving the list: its status, cause, group of its cause, times, the places that saw it, what our checks saw first and its activity log. The monitor's name in a row opens the monitor's page instead. The address in your browser keeps the open incident, so a bookmark or a link you send to someone in your account opens the same incident. Open the incident's page shows it in full. Download CSV, at the top of the page, gives the ongoing incidents and the list in one file.

The numbers and the chart

Above the list, four numbers sum up the time range, with the filters you chose:

  • Incidents: how many started in it (so it can differ from the list, which also holds those that started earlier).
  • Monitors affected: how many monitors had one.
  • Downtime: how long they were down in total, inside the time range.
  • Typical time to recover: how long an incident usually lasted until the monitor was up again (the middle value, so one very long outage does not hide the rest).

Under each, a line compares it with the period just before, of the same length, such as 4 more than the 7 days before.

Incidents over time shows when they started, a bar per hour for the last 24 hours and per day for longer ranges (a bar for each month beyond 90 days), each bar split by cause in the colours of its legend. Click a bar to show only that day, hour or month; choose another Time range to go back. Show the numbers as a table shows the same numbers, each period a link that does the same as its bar.

Filter the list

Search by a monitor's name or address. Add filter narrows the list further: choose Monitor, Group, Cause, Priority, Status or Place, then is or is not, then one or more values, and Apply. Each filter shows as a chip above the list, such as Cause is Timeout, DNS; click it to change it, or its cross to remove it. Clear filters removes them all.

  • Cause groups the causes: Timeout (no answer in time, from the website or to ping), Error status (the website answered with an error such as 500, 404 or 403, or showed an error page), DNS (the website's name was not found), SSL (a certificate problem, or no secure connection) and Other (anything else, such as a missing keyword, a closed port or a firewall's page). The panel shows each incident's group beside its cause.
  • Place is where our checks saw the monitor fail before the incident was confirmed.

Time range, at the right, chooses how far back the list goes: the Last 24 hours, 7 days, 30 days or 90 days, or all the history your plan shows. The address in your browser keeps the filters and the time range, so you can bookmark a view or send it to someone in your account.

From the Monitors list

In the Monitors list, a Down row has View incident: it opens the incident that is still going on in the panel at the side, without leaving the list of monitors.

An incident's page

Open the incident's page in the panel shows the incident in full: its status, cause, when it started, when two places confirmed it, when and how it ended, how long it lasted and the places that saw it. Go to monitor opens the monitor.

The Activity log tells the incident in order: each place's result, such as Amsterdam: Server error (502) and Newark confirmed it: Server error (502), each email it sent and when, and how it ended. Owners and Admins see the addresses the emails went to; everyone else sees how many. When a place saw the same failure several times in a row, those checks are one line, such as Newark: Server error (503), 6 times, with the time of the first and the last, so a long incident reads in a screen or two. When some lines are folded, the incident's page has Show every check, which shows each check on its own line, and Show fewer lines goes back.

Screenshots

For an HTTP(S) or Keyword monitor, the incident's page shows what the page looked like when it went down and when it came back, at each screen size we took, each with its time. Select a screenshot to open it at full size. Where one is missing, one line says why.

  • How it should look: beside the Down screenshot, the website's last good screenshot, when we still have one.
  • When we waited longer: when a site does not answer, we look once more and wait up to 135 seconds. If something comes back, such as Cloudflare's error 524, it is shown here.
  • A firewall's challenge is never shown.
  • Kept 30 days, like the screenshots in the alert emails.

The first failing answer

For an incident of a website, its page also shows First failing answer: what the website answered when it first failed, the status, the response headers and the first 2 KB of the page's visible text (its words, without its code). It helps you see what went wrong, such as an error page or a maintenance notice.

It is kept when the website itself failed, such as a server error or an error page. Some things are never kept: cookies and headers that may hold a secret, the page's code, anything from a website we check with a saved Basic Auth password, an incident caused by your own keyword or by statuses you chose, and anything from a port or ping monitor. The answer goes with its incident.

What changed before it broke

An incident's page, and its panel, list what changed on the website before it went down: a plugin or theme version, the web server, PHP, its DNS and your own account's changes. See What changed before it broke.

Download the list

Download CSV on the Incidents page gives every incident the page shows, with its search, filters and time range, in one file for a spreadsheet: the monitor, its address, the status, the cause, when it started and ended, how long it lasted in seconds, and the places. Its times are in UTC, so they read the same for everyone. It comes with every plan.

How far back incidents go

The page shows the months of history your plan shows, as the alerts do; an incident still going on is always shown. The longest Time range is that history, such as Last 3 months. With it chosen, and no filter, a line at the end of the list says how many months your plan shows when older incidents are kept but hidden.