Skip to content

Pause, snooze, acknowledge, disable

There are four ways to quiet something. They differ in what they quiet, whether the checking goes on, and whether your uptime changes.

Resolving an Incident by hand is not one of them: it ends the Incident and has the Check judged afresh. Incidents says how.

What it quiets Alerts Uptime Who may
Pause A whole Monitor: every Check stops. Open Incidents end, and nothing is sent until you resume. Paused time is left out, counted neither up nor down. Owners and admins
Snooze One Check, which keeps running. Messages about it wait until the snooze ends. Counted as usual. Everyone
Acknowledge One open Incident. Repeat alerts stop. The recovery message still comes. Counted as usual. Everyone
Disable One Check, which stops. Its open Incident ends, and nothing more is sent about it. Unchanged: the uptime percentage is the uptime Check’s, which is never turned off. Owners and admins

Members and Roles lists everything each Role may do.

Pause is in the Monitor’s header, and in the menu on its row in the list of Monitors. It stops every Check of the Monitor until somebody chooses Resume. The Monitor then reads “Paused” in place of its status.

When you pause, every open Incident on the Monitor ends. If you were told it was down, you get one notice: “acme.com was paused while it was down.” and “We will not check it until it is resumed.” If its alert had not gone out yet, nothing is sent at all. Anything about the Monitor still waiting to be sent is cancelled, and an open Regional outage ends in silence.

Paused time is left out of the uptime percentage entirely, so a planned shutdown is not downtime. On a Status page, a paused Monitor reads “Not monitored”. It still counts toward your Plan’s Monitors.

Resume starts afresh: every Check starts again from its first run, and nothing from the paused time is ever judged. A heartbeat Monitor’s URL still answers while it is paused, and a missed call opens nothing.

Snooze is on each Check’s row of the Checks card. Choose how long, or Until a time… to pick a moment: “The check keeps running. Messages about it wait until then.” A snooze ends at most 30 days away. The row then says until when, as “Snoozed until 14:00 UTC”, with End snooze to end it now.

While a Check is snoozed it runs and is judged as usual. Incidents open and end, and your uptime counts them. Only the messages wait. When the snooze ends:

  • An Incident that is still open sends its alert then.
  • An Incident that ended, and whose alert you had already had, sends its recovery message.
  • An Incident that began and ended while snoozed sends nothing, so a recovery is never the first thing you hear.

Any Check can be snoozed, the uptime Check too. A snoozed Check never joins a message about many Monitors. A Status page shows a snoozed Check exactly as a running one.

Acknowledge says “I am on it”. It is on the Incident card, on the Incident’s row on the Events tab, and behind the Acknowledge link in an alert.

It stops the Repeat alerts for that Incident at once. The recovery message still comes when it ends. The Incident then says who acknowledged it: “Acknowledged by Ada at 14:05 UTC”, or “from an alert” when it came through the link.

The link in an alert needs no sign-in. Opening it does not acknowledge anything by itself: it shows what you are acknowledging and one button to confirm, so a mail scanner that opens every link acknowledges nothing. It keeps working for 1 hour after the Incident ends.

When several Monitors went down together and came as one message, acknowledging it acknowledges them all.

Each Check’s row of the Checks card has a switch. Turning a Check off stops it, and the Monitor’s other Checks keep running.

If the Check had an open Incident, it ends. If you were told it was failing, you get one notice: “The certificate check on acme.com was turned off while it was failing.” Anything about it still waiting to be sent is cancelled.

The uptime Check cannot be turned off: “The uptime Check is always on.” To stop it, pause the Monitor. Turned back on, a Check starts again from its first run.