-
Notifications
You must be signed in to change notification settings - Fork 1.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
New CPU and memory usage metrics #888
Comments
I'm interest in this metrics! |
Another metric that might also be useful is the number of network connections that are currently open. And here's a library that could be used for getting some of those metrics in a cross-platform way: https://github.com/shirou/gopsutil/ |
It will also be useful to print a warning at the end of the test if the CPU/Mem has gone above a certain threshold. Possibly abort the test if we go above 99% memory used ? for example |
hmm yeah, having a default threshold (though probably not an aborting one) that checks we don't exceed some amount of RAM usage would be useful, so that users know the result of the test run would be skewed... |
If you are at 99% memory usage(probably at 95 as well) you are probably about to get the kernel to kill k6 at which point killing ourselves and reporting is probably MUCH better UX |
Digging for something unrelated, I stumbled upon https://golang.org/pkg/runtime/#ReadMemStats |
I am interested in this metrics! |
any updates on this? :) |
No, @poponuts. This has not been a priority. Additionally, I don't think we have agreed on any of the questions in the original comment, which likely isn't going to happen until this gets higher priority, sorry. |
Hi, FWIW this would be really useful for us at Alteryx. When running tests on Gitlab runners we have no other way of tracking the CPU and memory use of the tests, so we can't see if we need to increase the resources available to our runners or not. |
This seems like a good idea. if a load gen uses more than 80% cpu its going to start having issues. as stated, probably want to at least provide a warning, and offer an optional "abort if cpu>xx%" or something like that. memory is harder to determine (as if the kernel has a ton of RAM marked to be reclaimed, but the system shows it as used, we aren't going to have a problem, but if we have literally 95% actively used RAM that's different. So if the solution for RAM monitoring/alerting takes that into account that would be great. |
Maybe k6 could also measure the open file handlers ( |
It would be useful (especially after #570 is implemented) for k6 to measure and emit the used CPU and memory resources. Some open questions
The text was updated successfully, but these errors were encountered: