Azure Databricks resource in 30d free trial account takes FOREVER to DELETE. Also Databricks has a frequent SESSION EXPIRY problem.

Hi S V Dutt,
We are aware that several users are encountering the AADSTS90015 (Requested query string too long) error when attempting to sign in to Azure Databricks.

From the observed behavior—such as intermittent login success, impact across multiple tenants, and session drops—this indicates a service-side authentication issue involving Microsoft Entra ID and the Databricks sign-in flow, rather than a misconfiguration within individual workspaces.

As interim mitigation, you may try the following:

Access the workspace using its direct workspace URL instead of regional links or saved bookmarks

Clear browser cache and cookies, or use an Incognito/Private browsing session

Ensure only one account session is active during sign-in

Test with a different browser to rule out session-related inconsistencies

If the issue persists, collect the Request ID, Correlation ID, and Timestamp from the error page and share them when contacting support. These details will assist the service team in analyzing the backend authentication flow.

We are actively tracking similar occurrences and will provide updates as more information becomes available.

Thank you for your understanding.

Hi ,

Thanks for reaching out to Microsoft Q&A.

 You are hitting 2 diff but common issues with trial based setups.

1. Azure Databricks resource taking too long to delete

In a 30-day trial subscription on Microsoft Azure, Databricks deletion is slow because it is not a single resource. The workspace (adb-aw-project) has a managed resource group (managed-adb-aw-project) that contains VMs, disks, networking, and metadata. Azure first tears down all underlying compute and storage, which can take 30 minutes to several hours, and in some cases even longer if:

• clusters are still running or stuck

managed resource group has locked or orphaned resources

backend cleanup jobs are throttled in trial subscriptions

If it is stuck for many hours:

check if any clusters are still active and terminate them

try deleting the managed resource group manually

if still blocked, it is usually a backend issue → raise a support request (trial subscriptions are known to be slower)

2. Databricks session expiring every 5 minutes

This is not normal behaviour for Azure Databricks and is almost always caused by authentication/session handling issues, not Databricks itself.

Typical causes:

browser blocking third-party cookies (very common in Chrome/Edge strict mode)

corporate VPN or proxy resetting auth tokens

• AAD token refresh failure

system time mismatch (you already checked this)

What actually works:

allow third-party cookies for *.azuredatabricks.net and login.microsoftonline.com

try a different browser profile (not just incognito)

disable VPN temporarily and test

clear cookies specifically for Azure + Databricks domains

check if your org has conditional access policies forcing frequent re-auth

If it still happens every 5 minutes, it is almost certainly tenant-level Azure AD policy or browser restriction, not a Databricks bug.

Bottom line: Deletion delay = expected behaviour in trial + backend cleanup. Session expiry = browser/auth configuration issue, not platform instability.

Please 'Upvote'(Thumbs-up) and 'Accept' as answer if the reply was helpful. This will be benefitting other community members who face the same issue.