Increment of HDDs disk read request waiting time after linux kernel update

Issue presentation: - Environment: Debian (bullseye) based distro (proxmox VE) on top of ZFS - Action: Kernel update from 5.15.74-1 to 5.19.17-1 - Consequence: Increment of HDDs disk read request waiting time

I have a zabbix instance showing that the property "Disk read request avg waiting time (r_await)" has increased in a x2 - x3 factor after updating the kernel, from below 20 ms to more than 40 ms. In the machine I have 3 SSD, all of them unaffected; and 3 HDD, all of them affected by this issue in a similar way. Writing waiting time on HDD it's also unaffected (avg = 0.75 ms). If I run a read intensive process like if=/dev/sdx..., then the waiting time of the disk it's reduced to 2-3 ms. So the question are: - Why I have such high read request waiting time on all my HDD? It has increased after the kernel update, but before it was already way higher than HDD write waiting time (near 20 ms avg vs 0.75 ms avg) - Why it's reduced to 2-3 ms when doing intensive reading? Why it's still 3 times higher than write wait time? Best regards.