External Content Connector Confluence Cloud

[SuhejlA]

I have a question regarding a persistent timeout issue with the External Content Connector Atlassian Confluence Cloud.

Setup

• ServiceNow instance: regular company non-prod/TEST instance (not a PDI)
• App: External Content Connectors, sn_ext_conn v8.0.10 ("Up to date")
• Connector name: "Confluence Cloud TEST"
• Followed the official documentation "Configure crawl settings for an Atlassian Confluence Cloud external content connector" plus the related setup/prerequisite guide linked from the connector's "Configure Confluence Cloud → Read More" panel

Authentication setup

• OAuth 2.0 / 3LO app registered in the Atlassian Developer Console, Confluence API added
• Access type: Resource-level
• Grant Type in the ServiceNow Application Registry: initially "Resource Owner Password Credentials" (mismatch with the actual OAuth Entity Profile, which used "Authorization Code") corrected to "Authorization Code" to match; issue persisted afterward
• Scopes tested with Classic Scopes only (no Granular Scopes) to rule out a scope-type mix:
  • read:confluence-space.summary
  • read:confluence-content.summary
  • read:confluence-content.all
  • read:confluence-content.permission
  • read:confluence-user
  • read:confluence-groups
• Connection Settings additionally require a separate "API Key" field (alongside Client ID/Client Secret) — tested with both a classic (non-scoped) API token and a scoped API token, no difference in behavior
• OAuth consent was completed fully and is confirmed visible under Atlassian Account → Security → Connected Apps, showing the expected granted permissions (view content, attachments, spaces, groups, users, etc.)

Symptom

• "Save and validate" on the Connection Settings page succeeds without error.
• Every operation that requires a structural/listing response from Confluence fails after almost exactly 10 minutes with Job failed to start with the timeout. Job will be cancelled. / Crawl timeout after 10 minutes:
  • Start-Point Search in the "Crawl content" dialog ("No results found" when choosing a start point)
  • Crawl Type "Document (StartPoints Scan)"
  • Full Document Crawl
  • Partial Document Crawl
• The related "External Content Connector Crawl Events" log entry shows Successfully committed job to scheduler (Code: OTHER, Type: Error, Summary: "Crawl timeout after 10 minutes") confirming the job is created and queued correctly, but no further progress is logged before the timeout hits.

Already ruled out

• Inbound mTLS not enabled → confirmed active via /adcv2/supports_tls returning true, and plugin "Certificate-based authentication" (com.glide.auth.mutual) is active
• Incorrect Grant Type → corrected to Authorization Code
• Classic/Granular OAuth scope mix → issue persists with Classic Scopes only
• PDI hibernation behavior → this is a regular company TEST instance, not a Developer PDI
• Crawl schedule misconfiguration (frequency too high causing job pile-up) → schedules are Weekly; manual "Run now" fails identically and immediately shows the same timeout pattern
• Elevating to a higher-privilege role (ais_high_security_admin) in the session → no change in behavior or UI

Not verifiable from our side

• Whether the OAuth access token is actually persisted in oauth_credential this table sits in the protected application scope of sn_ext_conn (a Store app), so as customer admins we cannot execute scripts within that scope to confirm whether the token itself is present, even though the associated expires timestamp does update after each consent flow (consistent with a real token exchange occurring).

Has anyone encountered this exact pattern timing out after exactly 10 minutes with the Confluence Cloud connector?