-
Notifications
You must be signed in to change notification settings - Fork 30
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
Update performance benchmarks for v2 #163
Conversation
This now all works, just breaking up the changes into clearer PRs:
I'll rebase this PR purely with the benchmark updates after those are merged. |
I re-ran the benchmarks on a big VM, and the important thing is that with the changes here and in traefik v2, we no longer have an O(routes) problem, and performance is consistent and acceptable for up to 500 routes, which was the main thing we were after! #165 was critical for getting the benchmarks to run at all, and setting traefik.providersThrottleDuration to '0s' by default, we see performance much closer to CHP (still, CHP is closer to 1ms while traefik is closer to 100ms, but that's fast enough for us, where before it was closer to 2s because of the throttling) |
@GeorgianaElena Now that all the actual changes have been merged, this is purely an update to the benchmarks themselves, and a re-run of the data, updating the notebook. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think it'd be helpful to have a short performance/README.md
explaining where to start (probably just run <these-scripts>
)
Co-authored-by: Simon Li <[email protected]>
@manics good call. Added a README |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
🎉 🎉 🎉
I'm trying to get the benchmarks to run again so we can evaluate our current performance situation. But it's currently failing in most cases.
I've found a few performance things we can improve:
I'm still running into
cannot assign requested address
problems with consul, which appear to be failures in aiohttp? I know that's been an issue with aiohttp in the past, but haven't identified what can be done about it.placeholder PR for now while I'm making a mess, I'll add a comment when I think this is worth looking at.