Microsoft Fabric Trial Cap Issue Every Spark Notebook Immediately Disconnects & Capacity Admin Neve

Hi anjalik_06, 

Are you using the default spark pool? 

THe default pool is too large for the F4 capacity, so one spark session will cause the capacity to be over extended. 

I recommend creating a new pool that has a small node size and disabling auto scale.  

Hi anjalik_06,

Thank you for providing the detailed information. Based on what you've shared, you've already verified that the Fabric Trial is active, the required tenant settings are enabled, and the issue occurs across newly created workspaces and notebooks, which rules out many of the common configuration checks. Check the Microsoft guidance for Spark session startup issues. If the issue persists after completing the documented troubleshooting, the next step would be to open a Microsoft Support ticket and include the Diagnostic ID and Instance ID so the support team can investigate further.

Troubleshooting Guide for Spark Jobs in Microsoft Fabric - Microsoft Fabric | Microsoft Learn

Create a Fabric and Power BI Support Ticket - Power BI | Microsoft Learn

Thank you.

Hii sir 

Based on everything you've tested, this doesn't appear to be a notebook or workspace configuration issue.

The key indicators are:

• Spark sessions fail immediately for every notebook, including brand-new notebooks.

• The Trial Capacity is shown as Active, but the Fabric Admin field in Capacity Settings never finishes loading.

• The WebSocket closes almost immediately with:

  • code: 1000

  • state: "session-completed"

  • reason: "Session error or stopped."

• The issue is consistent across multiple workspaces and notebooks, with and without a Lakehouse attached.

• Microsoft Fabric Service Health reports no active incidents.

Since you've already ruled out the common causes (permissions, tenant settings, workspace configuration, environments, and Lakehouse attachment), this strongly suggests a trial capacity provisioning or backend compute issue rather than a problem with your notebooks.

At this point, I would recommend opening a Microsoft Support ticket and including:

• Diagnostic ID: 5079045f-6691-4a6d-9249-d50ef0b45b4a

• Instance ID: 5aed9132-d008-406b-acb6-d7050fa2f1ad

• Timestamp: 2026-07-10T00:38:44.223Z

• Capacity ID (visible in the WebSocket URL)

• Region: Central India

• Confirmation that the issue reproduces in newly created workspaces and notebooks.

The fact that the Fabric Admin section never loads is particularly significant, as it may indicate that the trial capacity was not provisioned correctly or is in an inconsistent state. This is something that Microsoft engineers can verify from the backend.

If this is a trial tenant and you don't have existing work to preserve, another workaround is to deactivate the current Fabric trial (if possible) and activate a new trial capacity. If the problem persists on a fresh trial, that would further confirm a backend provisioning issue.

Given the evidence you've collected, I don't believe there's any additional client-side troubleshooting to perform until Microsoft investigates the trial capacity itself.

hey anjalik_06, Strange indeed hmm, based on what you have shared, this looks more like a Fabric Trial Capacity provisioning/backend issue than a notebook issue, especially since the issue occurs across all notebooks and workspace, Spark sessions never successfully provision and the Trial capacity shows Active, but the Fabric Admin field never finishes loading.

Can you try:

• Opening a new browser / incognito session to rule out client-side issues?
• Creating a new Fabric trial capacity (if available) and testing again?
• Configuring a new Spark pool with smaller node size and disabling auto scale?

Appreciate if you can "Kudos" and/or "Accept as Solution" if this answered your query.