Suddenly performance of unraid terribly slow - especially visable on docker pulls

I have a very strange behaviour were suddently the performance on applications, NAS etc is all slower than where I started... Especially its visable when I update docker containers. This is a 10 Core xeon systems with 128GB memory but extracting is suddently ridiculusly slow...

All containers and vms should run on dedicated NVME m.2 ssds which perform great will directly tested...

When I check IO stat, it looks like the applications are reading very slow (disk idles most of the time and the writes very fast)

Any ideas whats happening here? I saw another tread about some plugins maybe interfeering but I would not know which one and some are quite important for me.

I have added the diagonistics and two videos showing the behavior

Intel® Xeon® CPU E5-2630 v4 @ 2.20GHz

SCSI Devices

[0:0:0:0]disk Kingston DT Workspace KS13 /dev/sda 64.0GB

[2:0:0:0]disk ATA SATADOM-ML 3MG-P 721 /dev/sdb 62.0GB

[5:0:0:0]disk IBM-E050 ST8000NM0065 BCE4 /dev/sdc 8.00TB

[5:0:1:0]disk IBM-E050 ST8000NM0065 BCE4 /dev/sdd 8.00TB

[5:0:2:0]disk IBM-E050 ST8000NM0065 BCE4 /dev/sde 8.00TB

[5:0:3:0]disk IBM-E050 ST8000NM0065 BCE4 /dev/sdf 8.00TB

[5:0:4:0]disk IBM-E050 ST8000NM0065 BCE4 /dev/sdg 8.00TB

[5:0:5:0]disk IBM-E050 ST8000NM0065 BCE4 /dev/sdh 8.00TB

[N:0:8215:1] disk WD Red SN700 250GB__1 /dev/nvme0n1 250GB

[N:1:8215:1] disk WD Red SN700 1000GB__1 /dev/nvme1n1 1.00TB

[N:2:8215:1] disk WD Red SN700 250GB__1 /dev/nvme2n1 250GB

[N:3:8215:1] disk WD Red SN700 1000GB__1 /dev/nvme3n1 1.00TB

the two 250GB ssds form a pool for caching the main raid

the two 1TB ssds for a pool for apps and vms only which shares are set to prefer this pool

All runs over one 10Gbit ethernet with multiple Vlans

unraid-diagnostics-20230214-0947.zip