On January 26 and 27, we experienced delays in the post-processing of dictations. Post-processing includes persisting audio in cloud storage (Amazon S3), logging the dictation transaction information, and decrementing the concurrent dictation count for the user performing the dictation.
As a result, some users received error messages indicating that they had reached their concurrent dictation limit even though they had finished the prior dictations and received the transcribed text.
In version 14.0, we will free-up the concurrent dictation count immediately after dictations complete, allowing the user to start the next dictation immediately. Post-processing will then proceed asynchronously without any direct link to dictation processing.