(21-02-2017, 05:46 AM)mrtn Wrote: I actually haven't thought about this approach. Yet, I have to object about RAM and CPU usage. I have a pretty good understanding of how much RAM and CPU should be in use at any time on my system (approximately). If I see my indicators display a lot more than they should, I know that something might cause trouble. And from there on, I can investigate and check.

Let's say you're doing random things, and then you notice you're using 60% RAM instead of your usual 25%. You will start investigating, and after spending roughly 30 minutes of running `free -m` 4 times, `ps auxf` 6 times, reading the `top` manpage to know how to sort by memory usage, googling the meaning of VSZ, RSS, you'll either find out that your RAM usage is back to normal, or that you left a browser open with a video, that's causing no harm anyway because you still have 3.2Gib of RAM available for running your terminal + weechat window.

And even if you manage to fill those 8gigs, the computer will start swapping on your SSD, and you will only notice because your status bar displays "RAM: 93%", not because things gets slower.

The only reason to care about your RAM/CPU usage is when you start noticing that your computer doesn't run at its full potential. And when it does (I'm speaking for me at least), I truly notice it.
My mouse start jumping 50 pixels away, windows refuse to move, applications take longer to start..
When it does, I just pop a terminal, make it full screen and start investigating. Last time it happened because I was running a system upgrade in the background, and compiling gcc was eating all the RAM (not even CPU because it's tied to 1 core).
By running `free -m`, I noticed my swap was disabled, so I reenabled it and was of to go, while memory was still ~90% usage.