A connection is an external place an alert can go. Set one up once in the monitor editor and any notify step in any flow can point at it. Credentials are encrypted at rest, and a Send test button confirms it works before you rely on it.
Chat destinations use an incoming webhook, so setup is mostly a matter of pasting a URL.
Post to a channel through an incoming webhook. The most common first destination, and usually the one that stays.
Post to a channel through a webhook. Popular with smaller teams running their own infrastructure.
Post to a channel through an incoming webhook, for teams already in the Microsoft stack.
Post to a space through an incoming webhook, the same way the other chat destinations work.
Send a message to a chat or channel through your own Telegram bot, using a bot token and chat ID.
Send routine notices to a general channel and real outages to an incident channel, decided on the canvas.
Channels inform a team. These destinations reach a person who is not looking at a screen.
Built in and always available. Send to your own account address or a team member, with placeholders in the message, an optional Acknowledge link, and the option to follow the on-call rota instead of a fixed recipient.
Text a number through your own Twilio connection, so you control the sending number and pay Twilio directly.
Call a number through Twilio and read the alert aloud, twice. The loudest option available, for the alerts that genuinely warrant it.
Push an alert to your phone and desktop through Pushover, which is a lighter alternative to SMS for many teams.
Publish to an ntfy topic on ntfy.sh or on your own server, which keeps the delivery path entirely under your control.
An email step can page whoever is on duty right now rather than a fixed address, so the rota does the routing.
Trigger an incident when the monitor goes down and resolve it automatically on recovery, so nothing is left open by hand.
Open a ticket with a summary you choose, for outages that need to enter your normal work tracking.
Open an issue in a repository with a title and body you choose, which suits teams who already run operations out of GitHub.
POST the alert to a URL of your own with an optional bearer token. The escape hatch for anything not on this list.
The connection decides where an alert goes. These settings decide how often it goes there.
The default. A notify step fires once and resets on recovery, so a long outage does not bury you in repeats.
Switch any notify step to fire on every matching check when you want a steady signal rather than a single message.
Every notify step has a cooldown, fifteen minutes by default, so a monitor bouncing down and up will not re-send inside that window.
Compose the message with placeholders such as flow, state, detail and duration, so the alert carries context rather than just a name.
If you have only one connection of the right type, a new notify block picks it for you instead of making you choose.
Put a noisy destination behind set hours and days while leaving the critical path open around the clock.
Paste a webhook URL, send a test, and draw a line to it from your first monitor.