Kaldi is not stopping after silence

I am using rhasspy 2.5.11 (in docker) and using Kaldi as speech to text service.

Everything is working fine except that it takes almost exact 10 seconds before AsrRecordingFinished and transcription of the spoken audio started.
So it looks like the silence detection is not working or not sensitive enough.

I have no idea how I can troubleshoot this and/or check the silence detection mechanism.

strange it is working now (I changed some settings back to what it was) with following silence settings: