Confluence Migration Failing Specific Error Details Not Showing in UI

Question

Hi Atlassian Experts,

I'm migrating from Confluence Version 7.19.17 from Server to Confluence Cloud and multiple spaces are failing. The issue is that when spaces fail, i'm not getting the specific error details mentioned in your troubleshooting documentation.

The troubleshooting page mentions specific errors like:
- 'Failed to execute export for query'
- 'Not enough space available'
- 'ERROR: duplicate key value violates unique constraint'

But in My case, i'm only seeing generic messages like:
- 'Uploaded 389 bytes of 37 KB. Open the troubleshooting page and locate a fix under the Migration errors section.'
- 'Migration workflow completed prematurely. Open the troubleshooting page and locate a fix under the Migration errors section.'
- 'Exporting database. Open the troubleshooting page and locate a fix under the Migration errors section.'

These messages tell me to open the troubleshooting page, but they do not show me which specific error from that page applies to my case. I cannot find the matching fix because the exact error code or message is not displayed.

Could you anyone please help me to identify:
1. What is the actual underlying error for each of our failed spaces?
2. Where can we find the detailed error logs?
3. Which fix from the troubleshooting page applies to our situation?

--- MY MIGRATION DETAILS ---
Source: Confluence Server 7.19.17
Destination: https://XXX.atlassian.net
Total Spaces: 456

--- FAILED SPACES ---
All Spaces are getting failed

Thank you,

Alexander Nilsson · Answer

Hello Fatima, thanks for sharing the log lines, they pin down what is going on. To your first question: yes, this confirms a proxy or firewall problem on the outbound path from your Confluence node to Atlassian Cloud. The smoking gun is the combination of ProxyTunnelEventListener connectFailed and SocketTimeoutException: connect timed out after roughly 5006 ms. That is CCMA asking your forward proxy to open an HTTP CONNECT tunnel to an Atlassian endpoint, and the proxy never completing the handshake. The 8 spaces that did migrate vs 448 that failed is a typical signature for this. The migration agent reuses connections where it can, so a few early calls slip through while the bulk of later space uploads run into the same blocked or throttled CONNECT path and time out. To your second question, the exact list of endpoints CCMA needs to reach is maintained by Atlassian here: https://support.atlassian.com/organization-administration/docs/ip-addresses-and-domains-for-atlassian-cloud-products/ Scroll to the section titled "For Confluence Cloud Migration Assistant communication". The endpoints you specifically want on the allowlist for CCMA are api.media.atlassian.com, api-private.atlassian.com, api.atlassian.com, migration.atlassian.com, migration-service.services.atlassian.com, id.atlassian.com, developer.atlassian.com, marketplace.atlassian.com, the S3 upload bucket under *.s3.us-west-2.amazonaws.com (specifically rps--prod-west2--migration-catalogue--migration-storage.s3.us-west-2.amazonaws.com if you cannot wildcard), and the URL of your destination cloud site. Same page also lists the egress IP ranges if your network team prefers IP allowlisting over hostname allowlisting, but hostname allowlisting is much less brittle since the IPs do change. To your third question: bumping proxy timeouts is not the fix here. A 5 second connect timeout is normal for an HTTP CONNECT handshake, and the docs do not expose a CCMA-side timeout knob you can tune. If your proxy cannot complete the CONNECT in 5 seconds for these hosts, it almost certainly cannot complete it in 30 either, because the underlying issue is that the proxy is refusing or dropping the tunnel, not that it is slow. Raising the timeout only delays the same error. One more thing worth checking on your side, given the partial success: if your forward proxy does any kind of TLS inspection or HTTP/2 to HTTP/1.1 downgrade on Atlassian domains, exclude the Atlassian first-party domains from inspection. The same page above explicitly calls this out as a source of degraded behaviour, and it matches the pattern of some spaces working while most fail. Practical next step: ask your network team to confirm CONNECT to migration-service.services.atlassian.com:443 works from the Confluence host through the proxy, for example with curl -v -x http://yourproxy:port https://migration-service.services.atlassian.com. If that one call fails the same way, you have your reproducer outside of CCMA and the network team can troubleshoot it without involving the migration tool at all. Hopefully all going successful through after this 🙂 Greetings, Alex

Fatima Hashmi_46ww3a · Answer

Hello Alex,

Thank you for the guidance. I downloaded the logs from the Confluence server and found the following errors repeatedly during the migration:

2026-05-15 07:49:57,733 ERROR [CCMA-MigrationTracker-pool-20-thread-4] [migration.agent.okhttp.ProxyTunnelEventListener] connectFailed *** PROXY TUNNEL CREATION FAILED ***
2026-05-15 07:49:57,734 ERROR [CCMA-MigrationTracker-pool-20-thread-4] [migration.agent.okhttp.ProxyTunnelEventListener] connectFailed Connection failure details:
2026-05-15 07:49:57,734 ERROR [CCMA-MigrationTracker-pool-20-thread-4] [migration.agent.okhttp.ProxyTunnelLoggingInterceptor] intercept === HTTP Request Failed ===
2026-05-15 07:49:57,734 ERROR [CCMA-MigrationTracker-pool-20-thread-4] [migration.agent.okhttp.ProxyTunnelLoggingInterceptor] intercept Duration before failure: 5006 ms
2026-05-15 07:49:57,734 ERROR [CCMA-MigrationTracker-pool-20-thread-4] [migration.agent.okhttp.ProxyTunnelLoggingInterceptor] intercept Exception type: SocketTimeoutException
2026-05-15 07:49:57,734 ERROR [CCMA-MigrationTracker-pool-20-thread-4] [migration.agent.okhttp.ProxyTunnelLoggingInterceptor] intercept Exception message: connect timed out
2026-05-15 07:49:57,734 ERROR [CCMA-MigrationTracker-pool-20-thread-4] [migration.agent.okhttp.ProxyTunnelLoggingInterceptor] intercept CONNECTION ERROR: Proxy connection issue

It looks like the migration assistant is failing while trying to establish the proxy tunnel / outbound connection to Atlassian Cloud.

Also, after reviewing the migration summary, I noticed that not all spaces failed:

* Migrated successfully:

* 6426 users

* 19 groups

* 8 spaces

* Failed:

* 448 spaces

So it seems the migration can partially complete, but most spaces are failing because of the connectivity/proxy timeout issue.

Could you please help me understand:

* Whether this confirms a proxy/firewall/network configuration problem?

* Which Atlassian endpoints should be whitelisted?

* Whether increasing proxy timeout values could help in this situation?

Thank you again for your help.

Alexander Nilsson · Answer

Hello Fatima, those generic messages are by design, not a bug. The CCMA UI never shows the real cause, it always points at the troubleshooting page. Atlassian acknowledges this gap in their own public issue tracker, ticket MIG-2440, currently with status "Gathering Interest", and the documented workaround there is exactly "check logs or contact support": https://jira.atlassian.com/browse/MIG-2440 The actual error is in atlassian-confluence.log on your Confluence Application Server. For every failed step the Migration Assistant writes a line in this shape: ERROR Step failed, taskName: Migrate space | stepType: CONFLUENCE_IMPORT | message: The detailed logs are at: /logs/atlassian-confluence.log Also worth a look is catalina.out in the Tomcat logs directory, because thread exhaustion and OOM errors sometimes only surface there. The cleanest way to capture everything in one shot is to generate a Support ZIP from Confluence administration, which packages logs, config, and diagnostics together. Steps are here: https://confluence.atlassian.com/support/create-a-support-zip-790796819.html I cannot tell you which fix on the troubleshooting page applies without seeing your actual log line, and the troubleshooting page itself only covers three patterns. Many other CCMA failure modes have their own dedicated KB articles. One observation: all 456 spaces failing in the same early phase ("Uploaded 389 bytes of 37 KB", "Exporting database…", workflow completing prematurely) almost always means one global root cause, not 456 separate problems. Common candidates documented in the KBs are database privilege gaps (the PROCESS privilege one on MySQL), thread or heap exhaustion, blocked outbound connectivity to Atlassian migration endpoints, and an outdated CCMA plugin version. Your logs will tell you which one. Confluence Server reached end of support in February 2024, so if the log message does not map to an existing KB article, opening a ticket with the Support ZIP attached is the right next step: https://support.atlassian.com/contact/#/ Greetings and and hopefully everything will go smoothly with the migration later on. Alex