Speech-to-Text batch transcribe API in germanycentralwest doesn't work

Last Friday (May 31 2024) we started getting the following errors on all transcripts sent to the batch transcription API on our speech resource in germanycentralwest:

{"error":{"code":"InvalidData","message":"The audio format is invalid or cannot be detected."}}

We didn't change anything in our setup, just the service stopped working for us. Using exactly the same code with a speech resource in a different location works without any problems.

Just for completeness' sake and because the Azure support on twitter tried to convince me that the issue is with the format of the audio files (which has been the same for almost three years and as I said works without any problems on speech resources that aren't on germanycentralwest) here's the ffmpeg command we use to pre-process the audio:

ffmpeg -i input.mp3 -ac 1 -ab 128000 -af aresample=async=1 -ar 16000 -f mp3 result.mp3

Speech in germanycentralwest not working has already happened once before on April 15 2024 and resolved itself after a few days. It's really annoying because we have customers that require us to process data in Germany.
Please fix this, I'd be happy to provide more information if needed.