Issue Summary

When attempting to configure the Jira cloud for Slack integration at the project level (via Project settings → Apps → Slack integration), creating or loading project connections fails with multiple "something went wrong" errors, and the Slack OAuth dialog never appears. This affects multiple projects and users (including Atlassian support staff) on at least one customer site where the global Slack integration is enabled, and project permissions are correctly configured. The Slack for Jira engineering team has confirmed that Slack-side rate limiting is in place and that further dedicated engineering work with Slack is required.

Steps to Reproduce

1. In a Jira Cloud project (for example, Software or JSM), go to Project settings → Apps → Slack integration

Expected Results

The Slack OAuth/permissions pop-up should appear so the user can authorize or confirm the connection.

Actual Results

• Immediately after accessing the page (before any OAuth or permissions pop-up), the UI shows errors such as:

• Slack error: something went wrong while creating project connection
  Error ID: <> (IDs vary per attempt)

• Slack integration error: something went wrong while getting the project connection list

• When Atlassian support reproduces the issue, additional errors are observed:

• • Jira error: something went wrong while getting teams list
    (Error ID: 6a2e29e0-0385-11f1-a27d-69921ec9061d)

• • Client-side error. Error ID: 3e703d7eeb394e4a92cf24d6530be4fc. Please reload the page or try again later.

• The Slack OAuth/permissions dialog never appears, and the project connection is not created.

• Behavior reproduces across multiple browsers and different Jira/Slack users, including Atlassian admins.

• Global Slack integration is turned on (Jira Cloud for Slack app enabled, workspace allowed), and atlassian-addons-project-access plus the configuring user both have Browse projects on the affected projects, so the issue is not resolved by the standard configuration steps.

Workaround

Currently, there is no known workaround for this behavior. A workaround will be added here when available.