My alert didn't fire

The sync succeeded, the change appears in the in-app inbox, but Slack stayed quiet. That's almost always a routing problem, not a delivery problem. Walk this list.

1. Confirm the change is real

Open the change in Inbox. If it's there, the diff engine did its job — the gap is between "change exists" and "Slack got pinged." That narrows things significantly.

2. Check severity routing

Settings → Alert preferences. Look at the matrix. Does the severity tier of the change you missed have a checkmark for the channel you expected?

A common pattern: someone restricts PagerDuty to Breaking, then forgets they also restricted Slack. Risky changes go to nobody.

Fix it: turn on the channel for the severity tier you care about.

3. Check per-integration overrides

Org-level routing is the default; integrations can override it. Integration → Alert preferences. If "Use org defaults" is off, the integration has its own rules — and they may be tighter than you remember.

If the override was set deliberately, make sure it's still right. If not, switch back to org defaults.

4. Check quiet hours

Settings → Alert preferences → Quiet hours. If quiet hours are configured, non-breaking alerts during the window get queued, not delivered. They show up after the window ends, batched.

Breaking changes always get through. If you missed a breaking change, quiet hours aren't the cause.

5. Check the channel itself

For Slack and Teams: did the channel get archived? Did the bot get removed? Did someone change the default channel?

Settings → Integrations → Slack shows the connected default. Reconnect if the install was revoked.

For PagerDuty: routing key still valid? PagerDuty service still active?

For email: addresses still valid? Check spam.

6. Check the re-ping schedule

If a change is Open and unacknowledged, it'll re-ping based on the re-ping schedule. If you only saw the original alert and missed it, the re-ping should follow.

Settings → Alert preferences → Re-ping for the schedule.

7. Check duplicate suppression

If twelve endpoints changed because of one underlying schema change, Intello sends one alert, not twelve. The other eleven exist in the inbox but didn't ping. That's correct behavior — you don't want twelve pings for one cause — but it can look like alerts are missing.

The change card in the inbox shows "+ N related changes" when this happens.