Alert channels
A channel is where alerts go. Channels belong to your organization, and any project's rules can send to them. Connect them under Settings → Alert channels, or from a rule while you write it.
Every organization starts with one channel, Organization owners, and every project with one rule, Any check down, sending to it. So the first check that goes down emails you without any setup. Owners are emailed at their verified address only — verify yours from the link in your sign-up email.
Connecting each kind
| Kind | What to paste | Where to find it |
|---|---|---|
| Slack | The incoming webhook URL | Create a Slack app with Incoming Webhooks, add one to the channel, copy its URL |
| Microsoft Teams | The Workflows webhook URL | In the channel, add the Workflows template Post to a channel when a webhook request is received, copy its URL |
| Discord | The webhook URL | Server settings → Integrations → Webhooks → New webhook → Copy webhook URL |
| PagerDuty | The integration key, and your region (US or EU) | On a service: Integrations → Add integration → Events API V2, copy the Integration Key |
| Webhook | Your endpoint's https:// URL | See Receiving webhooks. nexovue generates the signing secret and shows it once |
| One or more addresses | — |
A webhook URL for Slack, Teams or Discord is a password: anyone holding it can post to your channel. nexovue stores it encrypted, shows only its host, and never returns it. To change it, paste a new one.
Before you rely on it: send a test
Every channel's page has Send test. It sends a message labelled as a test through exactly the path a real alert takes, and tells you what happened — or why it did not, in the receiver's own words. A PagerDuty test is resolved the moment it is triggered, so nobody stays paged.
What a message says
The rule's name, how bad it is, what happened, why, when it started, how long it lasted once it is over, and one link to the incident's page in nexovue. That page says what happened and who was told, and links on to the evidence — the check and its runs, the service, or the error — for as long as it exists: a message about a check you have since deleted still opens the incident. Never a request or response body, a header, a secret, or a stack trace: those stay behind the link.
Keeping it quiet
- Grouping. When several checks fail within half a minute of each other, a channel gets one message listing them. Each still has its own incident.
- At most ten a minute. Past that, the rest of the minute arrives as one summary.
- Flapping. A check that goes down more than four times in an hour is said to be flapping, once; after that you hear nothing about it until it has stayed up for 30 minutes.
- Quiet hours (email, Slack, Teams, Discord). Warnings raised during them wait and arrive together when they end. Critical alerts are never held.
- Mute a channel or a rule for an hour, four hours, or until 09:00 tomorrow. What fires during a mute is recorded on the Alerts screen and never sent.
PagerDuty and webhooks are read by machines, so they get one event per incident, never grouped or held.
When a channel stops working
A failed delivery is retried for up to an hour — on a timeout, a network error, a
429 (honouring Retry-After) or a 5xx. Any other refusal is final. Every attempt
is in the channel's Delivery log. A channel that has delivered nothing for an
hour is marked failing, and the organization's owners are emailed once — by
email, never through the broken channel. The next successful delivery makes it
healthy again.