Poor performance in VM hosting Plex

[dvdwsn]

I have a Ubuntu Server VM on Proxmox 7.4-16 with Plex in a Docker container.
I also have TrueNAS Scale running as a VM on this same Proxmox host. TrueNAS has 2 pools it shares via NFS to the Ubuntu VM.
Other docker containers run fine, but they aren't as demanding as Plex. The CPU is set to Host, I've checked with egrep '(vmx|svm)' --color=always /proc/cpuinfo and the VM has vmx enabled.
I'm not seeing any indication of CPU, disk IO, network IO, etc, bottlenecks on the PVE host, on the VM summary or in TrueNAS.
The Plex server does sort of work, it's just painfully slow and flaky. Also this Docker container was previously running on OMV and worked great. This is the same CPU, mobo, etc that OMV ran on. There is an extra level of virtualization now, so I'm wondering if that's what's causing the issues?

 

[Dunuin]

How many vCPUs did you give your VM? Transcoding in software by the CPU without hardware acceleration by a GPU is veeeery inefficent. I for example gave my emby VM 8 of the 16 cores and it really needs all of those.

 

[dvdwsn]

I gave it 6 vCPUs. I wasn't seeing any CPU spikes in Plex or the VM, but I can bump it up and try.
I should add that the slow and flakiness is from navigating the UI and trying to play music. I can't get it to stream video at all.
My CPU does have an iGPU that does transcoding, but I haven't passed that through yet.