@Mikael_Ollars the % of committed memory remains around 10% in that one day snapshot which seems great for Roon.
With average committed memory normally in the 31-34% range, what is that in [mega|giga]bytes of RAM?
Thanks, but I use a Business ChatGPT subscription.
The rest of my AI usage is limited to Co-Pliot inside Office (on MacOS), occassionally Apple AI, Rovo instead Jira and Google AI inside Chrome.
At present, I am only interested in the physical RAM usage, to see whether my NUC needs an upgrade to from 16GB to 32GB - not that it is crashing, it has been ROCK-solid (if you excuse the pun) since 2016.
1 Like
I use an non-priviledged Podman container for Roon, which is one of many, on a host with 16 cores and 64 GB memory. So, I doubt I can get more data from Grafana.
Roon uses by far the most memory, but CPU is typically very low. I’ve not set any restrictions on CPU or memory as my box isn’t resource constrained.
1 Like
I must have rebooted around the time of my last post, for some maintenance reason.
Uptime registered today, pretty close to 6days:
Note that Roon is using less memory, by a gigabyte niw than the last snapshot.
Oh, my library backup size is about 7-8Gb, library of 250K tracks and some 15K albums spread over 11Tb.
2 Likes
It looks like 2.65 (build 1653) has freed up over 1 GB memory, although not much data yet.
.
Not happy about those CPU spikes. I have to bounce the container to restore normality.
2 Likes
I’d be interested to see a version of this graph that covers less time (maybe 1-2 weeks?) but has enough resolution on the time axis to see when during the day the spikes happen.
I also want to make sure, when you say CPU spikes here do you mean:
- the very short one that seems to occur once a day
- the section from 4/25 to 4/27 where the “floor” of the CPU usage went up to 100%?
1 Like
It’s not really a spike, rather a single core gets pegged at 100% for no apparent reason. The only resolution is to restart Roon Server.
It doesn’t happen often, i.e. 3 times since the beginning of March, and I have yet to see it with 2.65. However, this may explain why some users report sudden poor performance and mitigate this with frequent reboots.
My container has no CPU restictions, so I’m not seeing any impact, but I imagine a two or four core i3 would.
Backgound analysis settings:
I have the logs for this and earlier periods.
1 Like
Thank you 
it’s nice to see the background work scheduling changes actually working
It reminds me vaguely of a bug I fixed where we’d spin a single CPU core forever after some particular failure in tearing down a Sonos stream. I don’t think we’re going to fix something like that with generic performance work, it’s much more of a single bug fix type thing. If you notice some action you take or thing that happens at the same time as the CPU usage going up we’d be interested, but I know how hard that is for uncommon events like this.
(also, for everyone here, a disclaimer: I’m not actually part of the tech support team, reading community posts reliably is not one of my better skills, I may not see messages here)
3 Likes
I thought I’d post a screenshot of the past 90 days.
At the end of April, a fix for gardbage collection was released, and you can see Roon memory consumption fall. However, with the more recent releases, it is increasing significantly. The only positive is no reoccurence of a CPU pegged at 100%.
3 Likes