Integrations2 min read

Configure Slack service channels

Route service incident messages to one or more Slack destinations.

Before you begin

Connect the Slack workspace and ensure the bot can access each target channel.

Open the feature

Open Services → select the service → Notifications → Slack. Workspace connection and service routing are separate: this page chooses where that service sends incident messages.

Configure destinations

  1. Enable Slack delivery for the service.
  2. Select a connected workspace channel.
  3. Add up to three active channels for the service.
  4. Save the service configuration.
  5. Use Test for every destination.

All linked channels receive lifecycle messages; they are not a fallback chain. At the three-channel limit, unlink an obsolete channel before adding another. Channel destinations are independent from incident war rooms and user identity links.

What OpsKnight does

All active destinations receive lifecycle messages; they are not a priority or fallback chain. OpsKnight stores durable destination identities and updates the incident projection as state changes. Routine service channels are separate from incident war rooms and identity links.

Verify it worked

Trigger a synthetic incident. Confirm exactly one message in every mapped channel and verify that later lifecycle changes update the existing projection.

Remove or change a destination

Open the same service page, unlink the obsolete destination, save, and send tests to the remaining destinations. Removing a routine destination does not close an existing war room. At the three-channel limit, remove an obsolete channel before adding another.

Troubleshooting

Channel does not appear: confirm the bot can discover/access it and that required public/private scopes and membership are present.

Test succeeds but incidents do not arrive: check service notification routing and the incident's service, then inspect delivery history/provider result.

Duplicate messages: confirm the channel is not mapped through multiple active destinations and inspect the incident delivery/projection IDs before retrying.

Next steps

Last updated for v2.0.0

Edit this page on GitHub