This was caused by a faulted processing service on a single application node, which was removed, refreshed and re-introduced.
Technical teams are continuing to work towards improved ways to handle single-node issues with faster automated recovery and improved performance capabilities.