Downtime alerts: channels, testing and quiet periods

An alert nobody receives is worse than no alert at all, because you believe you are covered. This page is the whole of it: where downtime notifications go, how to add a channel, how to prove it works, and how to stay quiet while you are working on a site on purpose.

Your account e-mail is already on

From the moment a site is connected, downtime notifications go to the e-mail address of your account. Nothing has to be set up, and it works on every plan including Free. You are written to when an incident opens and again when the site recovers.

You can turn that off with Send alerts to my account email on Monitoring → Alert channels — sensible if you have added a chat channel and would rather not get both. With it off and no other channel, monitoring still records everything; it just stops telling you.

There is one e-mail address, and it is yours

You cannot add a second e-mail recipient. Alerts go to the address the account signs in with, and nowhere else.

That is a decision, not something missing: a free-text address in a settings field means sending mail to someone who never agreed to receive it, from an address they do not recognise, about a site they may not own. Sending to colleagues belongs with real team members who have accepted an invitation, and that is where it will arrive.

Telling several people today

Use a channel instead of a second address. From the Starter plan upwards you can add your own:

• Slack — paste the incoming webhook URL from your Slack app.
• Discord — paste the channel webhook URL.
• Telegram — a bot token and the chat it should write to, which can be a group.
• Custom webhook — your own endpoint, which has to be reachable over HTTPS on a public address. We POST a JSON body and send three headers with it: the event name, a timestamp, and an X-WPSafer-Signature computed as HMAC-SHA256 over “timestamp.body”, so your side can prove the call is ours. Redirects are not followed and we give up after five seconds. The signing secret is shown once, when the channel is created; New secret issues a new one and the old signature stops being accepted from that moment. The exact headers and body are printed on the Alert channels screen itself, under Custom webhook payload — you do not have to guess the shape from our side.

Everyone in that Slack channel, that Discord channel or that Telegram group gets the alert, which is what “notify the team” actually needs. Up to ten channels per account. A channel you want to keep but silence for a while can be switched off with Enable this channel rather than deleted.

There is no SMS channel and no phone call, on any plan. No page in the product claims one.

Test it before you rely on it

Each channel, and the account e-mail, has a Send test button. Press it and a test notification is delivered through exactly the path a real alert would take — same address, same signature, same formatting. Last delivery next to the channel shows how the previous send went.

Do this once per channel when you set it up. A webhook URL with a typo, a Slack app that was later removed or a Telegram bot that was never allowed into the group all look perfectly fine in the settings screen and deliver nothing. The test is one click, and it is the only thing that distinguishes a configured channel from a working one. You can repeat it once a minute per channel.

What arrives, and when it repeats

Two notifications per incident by default: one when it opens, one when the site comes back. Recovery is only sent to the channels that received the opening notification — telling a channel that a site it never heard about is fine again would be noise.

If you want to be nudged while a site stays down, pick an interval under Remind me while a site stays down: 30 minutes, an hour, two, six, twelve or a day. The default is Do not remind me.

Two limits work quietly in the background. A single channel takes at most 30 deliveries an hour; anything beyond that is postponed to the next round rather than dropped, so a bad hour across many sites cannot flood a chat room or get your webhook rate-limited. And a delivery that fails is retried up to three times before it is given up on.

Planned work: the maintenance window

Before you take a site down on purpose, open its Uptime tab and set a Maintenance window — a start and an end, at most seven days apart.

Inside the window, checks keep running and incidents are still recorded, so your history stays honest, but no alert is sent. An incident that opens and closes inside the window is never announced at all. If the site is still down when the window ends, the opening alert goes out then — the window delays the alert, it does not cancel it.

You can end a window early with Remove window. The site list and the Monitoring page both mark sites whose alerts are paused for maintenance, so a forgotten window is visible rather than silent.

Which plans

Account e-mail, reminders and maintenance windows are on every plan. Slack, Discord, Telegram and custom webhooks are included from Starter upwards. If you move to a lower plan your channels are not deleted — they simply stop delivering until the plan allows them again.

Alerts for things that are not downtime — a domain name about to expire, a DNS record that changed — follow the same channels and are described in how uptime monitoring works.