Sorry for overlooking that important detail. The main reason why I am using the official Docker container for Plex on my DS218+ is for automatic PMS updates (a simple container restart triggers the package upgrade).
These are my linkages :
(Synology DSM Host → Path in Plex container)
/volume1/Plex → /config
/volume1/Plex/tmp_transcoding → /transcode
Host : /dev/dri → /dev/dri (Container)
Running these chmod commands in the host Synology DSM on boot (automatic scheduled task) :
chmod 755 /dev/dri
chmod 666 /dev/dri/*
As a result inside my container, running the command ls -l /dev/dri returns :
total 0
crw-rw-rw- 1 root root 226, 0 Nov 19 14:47 card0
crw-rw-rw- 1 root root 226, 64 Nov 19 14:47 controlD64
crw-rw-rw- 1 root root 226, 128 Nov 19 14:47 renderD128
With all this, I get proper HW transcoding, with barely 0% CPU usage while transcoding a 1080p H264 stream with no audio transcoding (Direct Stream for audio, transcoding only for video).
As soon as the source video stream is interlaced, a fairly big amount of CPU usage gets added when transcoding, as shown in my previous post. 
And if you wish to know the full story, to fix the color corruption for HEVC 10 bit transcodes, I run the following command after every PMS upgrade :
mv /usr/lib/plexmediaserver/Plex\ Transcoder /usr/lib/plexmediaserver/Plex_Transcoder
and then creating the new Plex Transcoder file, containing :
cat /usr/lib/plexmediaserver/Plex\ Transcoder
#!/bin/bash
DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
declare -a new_args
forbidden_text=":format=nv12"
for param in "$@"; do
new_args+=("${param//$forbidden_text/}")
done
"${DIR}/Plex_Transcoder" "${new_args[@]}"
Hope you see it all clearly now 
If you could implement that last fix for J3355/J3455/N4200 intel cpus in all PMS builds, that would be great (until this gets definitely fixed by Intel)