Increasing "Error while fetching work item references" errors

[jmezach]

We are increasingly seeing errors such as those below in our CI pipelines:

Azure DevOps: Error while fetching work item references from Azure DevOps: The server returned a sign-in page. Please confirm the Personal Access Token is configured correctly in Azure DevOps Issue Tracker settings, and is authorized to read scopes 'Build', 'Work items' and 'Project and Team'.
Azure DevOps: Error while fetching work item references from Azure DevOps: The server returned a sign-in page. Please confirm the Personal Access Token is configured correctly in Azure DevOps Issue Tracker settings, and is authorized to read scopes 'Build', 'Work items' and 'Project and Team'.
Azure DevOps: Error while fetching work item references from Azure DevOps: 500 (InternalServerError). "Unable to complete authentication for user due to looping logins"
Azure DevOps: Error while fetching work item references from Azure DevOps: 500 (InternalServerError). "Unable to complete authentication for user due to looping logins"
Error: There was a problem with your request.
Azure DevOps: Error while fetching work item references from Azure DevOps: The server returned a sign-in page. Please confirm the Personal Access Token is configured correctly in Azure DevOps Issue Tracker settings, and is authorized to read scopes 'Build', 'Work items' and 'Project and Team'.
Azure DevOps: Error while fetching work item references from Azure DevOps: The server returned a sign-in page. Please confirm the Personal Access Token is configured correctly in Azure DevOps Issue Tracker settings, and is authorized to read scopes 'Build', 'Work items' and 'Project and Team'.
Azure DevOps: Error while fetching work item references from Azure DevOps: 500 (InternalServerError). "Unable to complete authentication for user due to looping logins"
Azure DevOps: Error while fetching work item references from Azure DevOps: 500 (InternalServerError). "Unable to complete authentication for user due to looping logins"
    at new t (D:\agent\_work\_tasks\OctopusCreateRelease_4e131b60-5532-4362-95b6-7c67d9841b4f\6.0.968\index.js:2:1890)
    at gX (D:\agent\_work\_tasks\OctopusCreateRelease_4e131b60-5532-4362-95b6-7c67d9841b4f\6.0.968\index.js:46:3493)
    at e.handleError (D:\agent\_work\_tasks\OctopusCreateRelease_4e131b60-5532-4362-95b6-7c67d9841b4f\6.0.968\index.js:46:2689)
    at e.<anonymous> (D:\agent\_work\_tasks\OctopusCreateRelease_4e131b60-5532-4362-95b6-7c67d9841b4f\6.0.968\index.js:46:3272)
    at u (D:\agent\_work\_tasks\OctopusCreateRelease_4e131b60-5532-4362-95b6-7c67d9841b4f\6.0.968\index.js:46:2146)
    at Object.throw (D:\agent\_work\_tasks\OctopusCreateRelease_4e131b60-5532-4362-95b6-7c67d9841b4f\6.0.968\index.js:46:1429)
    at u (D:\agent\_work\_tasks\OctopusCreateRelease_4e131b60-5532-4362-95b6-7c67d9841b4f\6.0.968\index.js:46:1038)
    at processTicksAndRejections (node:internal/process/task_queues:96:5)

We've verified that the token configured in Octopus Deploy is correct and using the Test button there also results in a success. It is not failing every time, but we are seeing it more and more often.