Status field is corrupted by a concurrent request to update an issue while a transition is in flight

If a transition is in progress for an issue and another request to update the issue occurs then the second update may rewrite the issue status to the value it had before the transition occurred.

This is happens intermittently as the result of a race condition to update the issue.

An example can be clearly seen in the changelog for GHS-4767, https://jira.atlassian.com/rest/api/2/issue/GHS-4767?expand=changelog at time 2012-05-10T18:27:30.416-0500 .