@darckhartyou're issue might be related to this:
https://forums.unraid.net/bug-reports/prereleases/690-beta-30-pre-skylake-intel-cpus-stuck-at-lowest-pstate-r1108/?do=findComment&comment=11255
Then this is mostly likely the bottleneck, you should get better speeds from SSD to SSD (an note that contrary to popular belief not all SSDs are capable of 500MB/s+, especial for sustained writes).
Those mitigation were already on v6.8.3, my thinking is that could be some new mitigation or change specific to this kernel, also not a bad idea to retest with v6.8 if possible to confirm it's really a beta issue.
There are constant ATA errors on disk1, that by itself shouldn't cause sync errors but if nothing else will make it much slower, replace cables and run another check, then post new diags.
Parity checks are single threaded, so one core stuck at 100% will bottleneck, but yeah, that CPU should be enough for the array size, unless like mentioned some kernel mitigation is making a large difference.
Diags show a high CPU utilization, suggesting that is the problem, or at least part of it, maybe the latest kernel vulnerability mitigations have an impact.
COW enable has some disadvantages for VMs, like increased fragmentation and write amplification, that's why LT disabled it by default for those shares, but I still prefer to have it enable for data integrity.