Confluence

The Confluence integration reads pages from the Atlassian spaces you connect and makes them retrievable in chat. Use it for design docs, ADR spaces, runbooks, onboarding wikis --- anything your team writes down that should be queryable alongside code.

Confluence and Jira share the same Atlassian OAuth app registration on your side. If you already set up Jira, most of this is done --- you just add Confluence-specific scopes and pick spaces.

Time budget: ~15 minutes.

What this connects

• Confluence pages in the spaces the connecting user has access to.
• Page hierarchy is preserved in the retrieval index so the platform can cite "under Space X → parent page Y → your page."
• Comments are not ingested in v1.
• Attachments are not ingested in v1 (linked docs from other integrations show up separately).

Confluence Cloud only. Confluence Server / Data Center (self-hosted) is not supported in v1.

Prerequisites

• Atlassian admin access on the Atlassian org that owns your Confluence workspace.
• Coordination with the platform operator.
• enable_confluence_integration = true in the tfvars.

Step 1: Register or reuse the Atlassian OAuth app

If you already have an Atlassian OAuth app registered for Jira, skip to Step 2 --- the same app works for both.

Otherwise:

1. Go to the Atlassian Developer Console.
2. Click Create → OAuth 2.0 integration.
3. Name it OutcomeOps AI Assist (or similar).
4. On the app page, note:
   • Client ID (top of the "Settings" tab)
   • Client Secret (in the "Settings" tab; regenerate if needed)

Step 2: Add Confluence scopes to the app

1. In the app, navigate to Permissions → Confluence API → Add.
2. Choose these scopes:
   • read:page:confluence
   • read:space:confluence
   • read:content:confluence
   • read:user:confluence
3. Save.

Step 3: Configure the OAuth callback URL

1. In the app, navigate to Authorization → OAuth 2.0 (3LO) → Add.
2. Callback URL: https://<your-app-fqdn>/api/confluence/callback.
3. Save.

Step 4: Handoff to the platform operator

The operator populates:

enable_confluence_integration = true
atlassian_oauth_client_id     = "<the Client ID>"

And creates the SSM SecureString for the client secret:

aws ssm put-parameter \
  --name "/prd/outcome-ops-ai-assist/oauth/atlassian/client-secret" \
  --value "<the Client Secret>" \
  --type SecureString \
  --key-id alias/aws/ssm \
  --region us-west-2

Note the SSM parameter path is /oauth/atlassian/ (shared with Jira), not Confluence-specific. If you already populated it for Jira, you don't populate it again.

Then terraform apply --- see the Deploy guide.

Step 5: Verify from the UI

1. Open the OutcomeOps UI as a workspace admin.
2. Navigate to Workspace Settings → Integrations.
3. Under Confluence, click Connect.
4. The browser opens Atlassian's consent screen. The user accepts.
5. The browser returns to Integrations with a green banner showing the connected user's display name.
6. Click Add spaces. The picker shows every Confluence space the connecting user can see.
7. Pick spaces and click Add. Sync status goes pending → in_progress → success.

Small spaces sync in a few minutes; large spaces (thousands of pages) take longer.

Common problems

Error	Cause	Fix
Redirect URI mismatch	Callback URL in the Atlassian app doesn't match the deploy's FQDN exactly.	Re-check the callback URL in the Developer Console. Must be https://<your-app-fqdn>/api/confluence/callback --- no trailing slash.
insufficient_scope after consent	The Confluence scopes weren't added in Step 2.	Add the four Confluence scopes, save, and try connecting again.
Space missing from picker	Connecting user doesn't have permission on the space.	Give the user permission in Confluence, then click Refresh on the picker.
Content is stale hours after edit	Confluence sync runs hourly.	Click Sync now for that space, or wait for the next hourly tick.

Disconnecting

Workspace Settings → Integrations → Disconnect Confluence. This removes the connection + the tokens; ingested space content is deleted from the workspace's KB.