Weird behaviour with LT 6.4 and pipelinePrewarming=true

[ovizii]

First, I am using https://github.com/Erikvl87/docker-languagetool.

Up until image: meyay/languagetool:6.3a-5 the parameter langtool_pipelinePrewarming=true was working fine, meaning when I started the container, CPU usage spiked for maybe 30 seconds then stopped and LT worked perfectly.

Starting with image: meyay/languagetool:6.4 when I start the container, CPU usage spikes and never stops and LT is unusable. Setting pipelinePrewarming=false and the container works just fine.

Can someone use plain LT, not with this container please let me know if their LT 6.4 works with pipelinePrewarming=true?