Issue with Plex and GPU passthrough (after sometime, the docker isn't using the GPU anymore)

[majo23]

HI:)

About my NAS:
Im using the lasted Unraid version and this Plex Docker . I have a 1050 ti passthrough to the plexdocker (also installed the Nvidia Driver Plugin). The following Problem also existed with the last Versions of UnraidOS, nvidia driver and Plex Versions - so the issue seems to be unrelated to a version of one of them.(My NAS is running 24/7)

ASUS P8B-M + Intel® Xeon® CPU E3-1240 V2 + ECC RAM + GTX 1050 ti

I passedthrough the GPU by using the "--runtime=nvidia" parameter and the "Nvidia Visible Devices:" statement with the GPU ID (i guess u know what i mean^^ sry for my poor english)

 

 

Description of the problem:

Sometimes the videoquality in Plex is very poor. If that happens I cant select an other Playbackquality in the plexplayer. That is very strange, because normally i can choose the videoquality when playing a movie. 

 

ALSO im using the GPU Statistics plugin, and when this issue is occurring, it isn't showing any load on the GPU.... normally when playing a video, the plugin shows a bit of load on the GPU and it also shows that the plexdocker is using the GPU at the moment. 

So... i think there is a issue with the docker using the GPU... maybe the GPU is "disconnecting" from the docker, then the docker is faulty and shows bad quality because the rest quality would normally been provided by the GPU - but is then lacking....

 

As I said, when restarting the docker - all is fine again...

ANY IDEAS?
Any ways to get a bit more from the logs?