Though I can deactivate the swap partition from the HDD and afterwards it's running acceptable. Perhaps it's using the faster zram, but I don't remember noticing zram usage increasing as shown my swapon -s. Or am I missing something?

▲ blueflow 3 days ago | parent [-]

Do you have the pressure stall information data in /proc/pressure? If you have it enabled, reboot, use chrome until your get your stall again, and then look which of the files in there has the highest total= number, thats likely it.

I doubt that disabling swap reduces your stalls, but i'd like to see the numbers for that.

▲ ciupicri 2 days ago | parent [-]

I haven't rebooted yet, but Chrome went downhill again.

    # cat /proc/pressure/memory 
    some avg10=3.34 avg60=6.40 avg300=3.22 total=65245554
    full avg10=3.18 avg60=6.14 avg300=3.10 total=64646112

    # free -m
          total   used   free  shared  buff/cache  available
    Mem:  15874   8099    309     370        8162       7775
    Swap: 25599  10873  14726

Swap is composed from 4+5 GBs of zRAM (lzo&zstd) and 16 GBs on HDD. There's also some background I/O from time to time caused by torrent seeding. vm.swappiness = 13.

I misread your comment, so I didn't check all the files, but here's how /proc/pressure/ looks now after managing to close Chrome:

    cpu     some avg10=0.00 avg60=0.00 avg300=0.07 total=188606342
    cpu     full avg10=0.00 avg60=0.00 avg300=0.00 total=0
    io      some avg10=2.07 avg60=1.50 avg300=9.23 total=2149496347
    io      full avg10=2.07 avg60=1.47 avg300=8.79 total=2045347731
    irq     full avg10=0.11 avg60=0.20 avg300=0.23 total=288863055
    memory  some avg10=0.00 avg60=0.00 avg300=1.22 total=94771913
    memory  full avg10=0.00 avg60=0.00 avg300=1.18 total=93379177

	▲	blueflow 2 days ago \| parent [-]
		Okay this is a legit case where having swap on a slow HDD is the main reason for stalls. 10 GB swapped, $expletive.