We should get rid of average CPU utilization
theocharis.dev
We should get rid of average CPU utilization
1–10 of 25 posts
Re: We should get rid of average CPU utilization
#2great article thanks
Re: We should get rid of average CPU utilization
#3No, we shouldn't. We should measure latency if we care about latency.
Re: We should get rid of average CPU utilization
#4Same thing when it comes to memory. The rabbit hole goes on forever, and metrics lie to you if you don't know how to interpret them properly.
Re: We should get rid of average CPU utilization
#5TLDR; if app slow, give more resources
Re: We should get rid of average CPU utilization
#6Lovely read, if you’ve ever had even remotely similar issues (you think you’re looking at the right places but you’re not) it read like a detective novel.
Re: We should get rid of average CPU utilization
#7No, we shouldn't. We should measure latency if we care about latency.
I’ve come to realise that “wide logs” like OpenTelemetry traces are the only way to go, despite the expense of collecting and storing them with current technology.
As open source columnar databases improve, the cost will drop.
Re: We should get rid of average CPU utilization
#8TLDR; if app slow, give more resources
Or just don't put CPU limits: https://home.robusta.dev/blog/stop-using-cpu-limits
Re: We should get rid of average CPU utilization
#9It's well known that many throttling implementations are broken, usually by design. You shouldn't blame the CPU utilization metric for that footgun.
In a well designed scheduler, a task that has been granted an allotment of at least n cores, should never get throttled to less than n cores at any time. It can be limited to less than n cores if CPU utilization is at 100% and another task gets scheduled at the time, since that's unavoidable when you oversubscribe the available resources.
Re: We should get rid of average CPU utilization
#10TLDR; if app slow, give more resources
Writing better code is of course out of the question.