On the other hand, merely posting that because you read some people having a specific problem does not mean that the problem is what you say it is either. It is indeed an issue or problem perhaps. In this very thread, someone who migrated to Scale from Core had a higher CPU usage, which you then immediately assumed was k3s (could have been, or may not have been). Maybe the migration has a bug, maybe if he did a new install it wouldn't. In that case, the actual prolblem is the migration right? I can cherry pick problems people are having on freeBSD and state that freeBSD has some issue too as a few users had some similar sounding problem. That wouldn't make it valid. I can find thousands of issues posted here and make broad claims. That is what I feel you are doing here. I have heard few people complain of that, the last time someone did they found out that it was one of their apps, just last week, the thread was titled something about high CPU usage. Problem was solved. High cpu usage can be all sort of things, you want to associate that with Scale and k3s. But yes, indeed, having one more layer is going to add something to resource usage, we all know that, and that will be worse on lower powered machines. Undoubtedly. But Windows has more resource usage than dos too. As things progress, more power is needed but it doesn't make it a problem per se.
The linked to blog post is light on details. The github post is about small (embedded) systems by the description, and many of the linked to cases are as well. Not very compelling except if your point is k3s will take more resources than no k3s, well, yes, that's true of any software. Stating things like high load average is conflating what load average is vs cpu percent, etc. Note the issue is not reproduceable indicating it's something on the specific machine too. The github issue is full of Rapsberry Pi users (I use one) and that is not surprising at all. Yes, k3s undoubtedly uses > 0 resources and may be too much for a Pi. Note the very last post on github, the guy discovers it was a specific helm chart. It's hard to say what the actual problem is for all the other folks. The most compelling of your links was actually to a truenas user with lots of k3s processes. That was clearly a bug of some sort with Scale.
In any case, what this is about is someone is having some sort of problem with cpu usage. What that is is unknown except for the solved cases on github that claimed this and turned out to be something else. It could be k3s, but then you would expect far more posts about that in my view here on the truenas forums which I've seen very few of (yes, more than 0). And on github, people will chime in and say they have the "same problem" when it is totally different, this is not surprising or unusual. High CPU by itself means nothing as far as determining why.
Yes, I could also find a process on Core/freeBSD that does not exist on Scale and make some claim about how that problem would never appear on Scale, but that's stretching things. I used to use freeBSD and found it quite good actually.
I realize there is a lot of hatred (or bias against) here for/against Scale/Linux and I don't want to discuss that but many posts in these forums have that. Core is a great product and undoubtedly it's still more reliable, I believe we can agree on that. That will change over time. When you say things like "this is why I will never install Scale", Scale is Beta, etc. that kind of indicates where comments are coming from. When I read things like that, my first natural impulse is to think I will reduce the value of the entire posters comment due to bias. That's the natural inclination, but then I realize I shouldn't really.
Really to me, Scale is more complex to use than Core. I don't mind that as a retired It guy as tech is trivial for me, and thus my Scale machine runs perfectly with no issues really. But so many home users who have limited docker / Linux knowledge is definitely going to create a lot of posts with various issues. That will take a while yet to prevent.
I'm glad you are here volunteering your time, you have a lot of good posts and knowledge to bring to the community in general. I just see this differently but appreciate your view. I just wish there was less of Core vs Scale in these forums. And that's what I saw your comments as. I could well be wrong about that. But pretty sure I'll get a lot of blowback on this post!
)
