If you think hardware is the problem, read again the thread 
And if you think PEBKAC is the problem, again read the thread and learn docker. Just quoting you here:
Doing things with k8s, docker, vms and backbone networks is my daily job, so start by asking questions before judging.
Birds of a feather. It’s the inherent sarcasm, lack of communication, uneducated and arguably useless responses, and complete disconnect from his own internal departments that’s unacceptable. Some he has control over, some he doesn’t. Look at the sheer number of responses asking Plex for some official action or response on this issue. Instead of acknowledging the fact that this many people are being effectively ignored, you instead decide to spend your time defending indefensible employee responses and video memes? Way to go.
How about start supporting everybody and their brother before you judge me?
I’ve been doing this ■■■■ a lot longer than you think.
As for Docker, why use a chroot when it’s not really needed? Sure, when done correctly, it has advantages. Moving containers between the same vendor’s implementation is trivial but I’ve yet to find it where I can move from NAS to desktop/workstation/server and have it work. Just doesn’t happen because everyone does it differently.
My point about docker is: Most do it “Just because they can” then complain when it doesn’t work. Further, they don’t understand how it works.
You are the exception. Thank you for that refreshing voice. I stand corrected.
Glad you approve.
The post count is irrelevant.
The same goal could be achieved by a gentle bump tagging in the respective Plex member evenif they have to become more frequent.
Not a sudden tag accompanied by a tirade.
I am willing to try and help figure out where this is IF everyone is done throwing stones?
I am part time help. Period. I deal with a lot every day.
I am currently writing all new packaging (installer) for DEB files which will soon include RPM.
I also support Synology and QNAP (I wrote those installers too).
I’ve been writing software for 35+ years at the kernel and system application level.
I do know my way around but am not as quick as I was when 30.
We all in sync now?
Can we start fresh?
My initial reply in this thread was explaining why Subtitles use a single thread. - Easy
We then advanced to behavior changes from 1.14.1 → 1.15.x , I acknowledged this is known and they (Engineering and Build teams) are trying to work it out.
I acknowleged:
Is the above not enough of an acknowledgement of the problem?
Later, we had a bit off-topic introduced
I ask:
We then advance forward to where my report was accurate (NO NEWS) but also added UPNP accidentally (which was an equally important and concurrent problem) which is what seems to unravel this thread.
I later add (May 15)
Given what I have reported, and I have all the bug reports submitted,
I don’t know what else I can say or do to appease everyone.
If anyone wants greater visibility, Write to plex directly at www.plex.tv
let them know how you feel.
There is only so much we can do from the customer support side.
So where do we stand please?
@ChuckPa I think the issue is not that you weren’t clear, but that you (as customer support) are fairly disconnected from the inner-workings of the business. For example, where I work I sit not even 2 rows of cubicles away from customer support (I’m a software engineer as well). So I just find that disappointing and frustrating, if nothing else. And no frustration directed towards you, just Plex in general I guess.
Having said that, I don’t see any other channel to communicate about this. You said to go to plex.tv. At the bottom, I click Contact, I get this:
So it seems all roads lead to customer support. I mean, who else is going to look at this and supposedly have more insight, knowledge, and control over the situation than you will? The high CPU usage should be the only issue discussed here. Others may have shifted the topic, and I have no idea how UPNP or Docker is relevant, but we should focus on that. And ultimately find a way to get Plex to give us a roadmap and/or more visibility into the priority of that issue, whether it’s through you or some other means.
I just looked. I guess they did change that page and you have no choice.
YOu can take another option and run it that way. That’s your choice.
I can tell you now, because we in support have asked and were told no.
Plex Engineering does not give RoadMaps. Period. Never have, and likely never will.
We see new features maybe 1-2 days before you do. Sometimes you see them first in “Plex Pass Beta”.
To help with my sanity, I would greatly appreciate we keep one-topic per thread.
We can link them here as the focal point but to intertwine multiple discussions invites madness and mistakes which is what sparked the above misunderstandings.
To futher avoid any misunderstanding about my skill set,
Hi @ChuckPa my apologies, I’d totally forgotten about that thread!! Anyway, keep doing the good work… I’m fighting a EPG issue with others in another forum post… Grapefruit seems to be the issue me thinks, and I’m looking forward to getting the UK EPG fixed soon?? Need cartoon channels back for the kids…
Dave
I have heard, but not yet seen, what will become of the EPG issue.
I’ve only heard whispers but have seen nothing in working code.
I just want to say that I reverted back to 1.14.1.5488 and my server has been running better than it has in the past 5-6 months. Plex had been acting extra weird the last few weeks with CPU (4690k OC to 4.0) usage jumping to 90-100% with just me on the server. It completely locked the computer to the point where i couldn’t log in remotely to end the PMS process. I had to go do a hard reset which led to a bunch of other problems non plex-related.
So if anyone is having the issue, just revert back to the 1.14.1.5488 build.
For what it’s worth, I just tried out the latest version of PMS, 1.16.1.1246, but the issue still remains. I’m running a docker version of PMS with 12c/24t, but only some of the cores assigned to Plex. I did more closely observe what happens since none of the clients I normally use have this problem.
I played a movie file using an Xbox One X client, making sure both video and audio are direct streaming, and had SRT subtitles enabled. This caused the SRT subtitles to be transcoded to ASS and the container from MKV to MP4. This is expected, as for whatever reason the Xbox can’t direct play SRT subtitles. I immediately noticed that one thread maxed out at 100%, but as the movie played over a couple of minutes, the CPU usage continued to increase, making more threads go gradually to 100%. I hit 5 threads at 100% before giving up on the exercise, but from past experience this would eventually max out all the threads assigned to Plex.
Going back to 1.14.1.5488 fixed this issue. I hope this describes my issue well enough, and if it didn’t belong to the issue in this thread, I would be glad to open up another.
Everyday I enter my PMS. 
New release notes show:
The server could use large amounts of CPU when running on a network with certain routers with UPnP enabled (#9610)
I know that has been brought up here, not sure if it directly impacts subtitle transcoding. But thought I’d bring it up for someone to test.
I had some hope reading the changelog, but still the same problem…
So still no fix?
Still having issues with latest version. Cpu is not used so much when using chromecast or android app - it does only replicate on web client 
it really hurts my cpu, as temperatures rise for cores when they are used 100% one core at a time.
I’m starting to think to start giving my money to someone else, not plex as we don’t get any hope for things to be fixed here
I just caught other pex user using subtitles and asket to turn it off. In result server did run 10 °C cooler, used 16 watts less power, and cpu did run 10% instead of 30%.
Seems like a lot of overhead just for subtitles.
Also I tested that burned in subtitles help, but other user did not had such option for some reason.
upnp us turned off in my router, I run docker inside Unraid on Dell T30. Do you need more info to get this fixed?
Do you think running VM would help, as you mentioned it could be docker issue?
I don’t think it would help…
I have the same problem with a VM (KVM / oVirt)
You guys are awesome! Saved me a lot of headache! I thought I was the only one facing this problem!
I’m using Windows 10 Pro PC (not using docker or anything, just purely windows installation of Plex)
And I can confirm the main cause of this and can re-produce this exact problem.
The main culprit here is the ASS subtitle, PMS has a memory leak on CPU from ASS subtitle for me anyways, I’ve posted my problem here.
The one’s that got affected are:
If you let Plex Web App convert your audio from 5.1 to stereo, it will also convert your SRT subtitle to ASS (causing CPU Leak).
If you also play direct stream with audio 5.1, plex will also convert your SRT to ASS (causing CPU Leak)
If you use Plex Web App by choosing audio stereo (if your media has stereo and NOT let plex web app convert it) it will play SRT subtitle successfully, and hence, not cpu leak.
ASS Subtitles causes the CPU leak, and will slowly rise, up to its max peak until you terminate it, or when the media is ended.
A temporary fix if you do not have stereo as audio in your media or you want to play audio with 6 channels or more (5.1 or more) is to use plex media player that can handle this without converting SRT to ASS.
But you prefer to use the web app, then change your subtitles to always burn.
Other players I’ve personally tested that are not affected by this and can successfully play SRT subtitle with 5.1 audio without converting to ASS are:
I have also reverted to an older PMS version to 1.14, as the recent versions of 1.16 still has the cpu memory leak of using ASS subtitle.
TL;DR:
ASS encoded Subtitles are the main culprit to CPU leak for the recent versions of PMS, version 1.14.1.5488 (the last 1.14 version of PMS before 1.15) handles ASS subtitle well without any leaks.
3 months now 
