Unable to start Data Flow Debug session in Synapse Workspace - Debug session is already terminated.

Haitham 0 Reputation points
2026-02-15T10:27:28.0766667+00:00

Hello, I am using an 'Azure for Students' subscription and I'm facing a persistent issue with Azure Synapse Data Flows. Every time I try to start a Debug session, it terminates immediately with the error: "Debug session is already terminated."

Troubleshooting steps already taken:

Created a custom Integration Runtime (IR) in West Europe with 4 (+4 Driver) cores (General Purpose).

Set the Time to Live (TTL) to 10 and 30 minutes.

Verified that my Managed Identity (Object ID: Masking PII Information) has 'Storage Blob Data Contributor' role on the ADLS Gen2 account.

Test connection for the Linked Service is successful.

The issue seems to be related to Core Quota or environment provisioning for student subscriptions. Could you please check if there is a core limit blocking the Spark clusters from starting in my region?

Azure Analysis Services
0 comments

2 answers

Sort by: Oldest
Most helpful Newest Oldest

  1. AI answer

    2026-02-15T10:27:39.74+00:00

    The error message "Debug session is already terminated" when trying to start a Data Flow Debug session in Azure Synapse can be caused by several factors. Here are some troubleshooting steps you can consider:

    1. Check Core Quota: Since you are using an 'Azure for Students' subscription, there may be limitations on the number of cores available for your subscription type. You can check your current core quota and see if it is being exceeded.
    2. Integration Runtime Configuration: Ensure that your custom Integration Runtime (IR) is properly configured and is not experiencing issues. Sometimes, creating a new IR or adjusting its settings can resolve such issues.
    3. Time to Live (TTL): The TTL settings you have configured (10 and 30 minutes) should generally be sufficient, but you may want to experiment with different TTL values to see if it affects the behavior of the debug session.
    4. Region Provisioning: Make sure that the resources you are trying to access are provisioned in the same region as your Integration Runtime. Mismatched regions can lead to issues with starting sessions.
    5. Managed Identity Permissions: Although you have verified that your Managed Identity has the 'Storage Blob Data Contributor' role, ensure that it has the necessary permissions for all resources involved in the data flow.

    If the issue persists, consider reaching out to Azure support for further assistance, especially regarding core limits specific to student subscriptions.

    References:

    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

  2. Smaran Thoomu 35,870 Reputation points Microsoft External Staff Moderator
    2026-02-16T02:49:01.24+00:00

    Hi Haitham
    Thanks for explaining the steps you already tried - that helps.

    The error “Debug session is already terminated” usually means the Spark cluster for Data Flow debug could not start or was shut down right away.

    Since you’re using an Azure for Students subscription, the most common cause is compute/core limits. Student subscriptions often have lower quotas, and Data Flow debug needs Spark clusters, which require cores to spin up. If the quota isn’t enough, the session can fail immediately like this.

    A few things to check:

    • Go to your subscription → Usage + quotas and check your vCore quota for the region (West Europe).
    • Try a smaller Data Flow debug cluster size if possible.
    • Make sure your Synapse workspace and IR are in the same region.

    If your quota is low or zero for that region, you may need to request a quota increase. Some student subscriptions cannot increase quotas, in which case this becomes a subscription limitation.

    Since your storage access and linked service tests are working, this doesn’t look like a permissions issue.

    If quota looks fine and it still fails, then it would be best to open a support ticket so the backend logs can be checked.

    Hope this helps point you in the right direction.

    Was this answer helpful?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.