Skip to content
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

[feature] Inherit resource limits from cgroups #1336

Merged
merged 1 commit into from
Jan 17, 2023

Conversation

daenney
Copy link
Member

@daenney daenney commented Jan 13, 2023

Description

When GTS is running in a container runtime which has configured CPU or memory limits or under an init system that uses cgroups to impose CPU and memory limits the values the Go runtime sees for GOMAXPROCS and GOMEMLIMIT are still based on the host resources, not the cgroup (if the cgroup has such limits set).

At least for the throttling middlewares which use GOMAXPROCS to configure their queue size, this can result in GTS running with values too big compared to the resources that will actuall be available to it.

This introduces 2 dependencies which can pick up resource contraints from the current cgroup and tune the Go runtime accordingly. This should result in the different queues being appropriately sized and in general more predictable performance. These dependencies are a no-op on non-Linux systems or if running in a cgroup that doesn't set a limit on CPU or memory.

The automatic tuning of GOMEMLIMIT can be disabled by either explicitly setting GOMEMLIMIT yourself or by setting AUTOMEMLIMIT=off. The automatic tuning of GOMAXPROCS can similarly be counteracted by setting GOMAXPROCS yourself.

Checklist

Please put an x inside each checkbox to indicate that you've read and followed it: [ ] -> [x]

If this is a documentation change, only the first checkbox must be filled (you can delete the others if you want).

  • I/we have read the GoToSocial contribution guidelines.
  • I/we have discussed the proposed changes already, either in an issue on the repository, or in the Matrix chat.
  • I/we have performed a self-review of added code.
  • I/we have written code that is legible and maintainable by others.
  • I/we have commented the added code, particularly in hard-to-understand areas.
  • I/we have made any necessary changes to documentation.
  • I/we have added tests that cover new code.
  • I/we have run tests and they pass locally with the changes.
  • I/we have run go fmt ./... and golangci-lint run.

@daenney
Copy link
Member Author

daenney commented Jan 13, 2023

I added this in cmd/gotosocial/action/server since it's only the long running server process that needs this. But it could be moved into main if that's preferable.

When GTS is running in a container runtime which has configured CPU or
memory limits or under an init system that uses cgroups to impose CPU
and memory limits the values the Go runtime sees for GOMAXPROCS and
GOMEMLIMIT are still based on the host resources, not the cgroup.

At least for the throttling middlewares which use GOMAXPROCS to
configure their queue size, this can result in GTS running with values
too big compared to the resources that will actuall be available to it.

This introduces 2 dependencies which can pick up resource contraints
from the current cgroup and tune the Go runtime accordingly. This should
result in the different queues being appropriately sized and in general
more predictable performance. These dependencies are a no-op on
non-Linux systems or if running in a cgroup that doesn't set a limit on
CPU or memory.

The automatic tuning of GOMEMLIMIT can be disabled by either explicitly
setting GOMEMLIMIT yourself or by setting AUTOMEMLIMIT=off. The
automatic tuning of GOMAXPROCS can similarly be counteracted by setting
GOMAXPROCS yourself.
@NyaaaWhatsUpDoc NyaaaWhatsUpDoc merged commit acc333c into superseriousbusiness:main Jan 17, 2023
@daenney daenney deleted the usage-limits branch January 17, 2023 20:59
daenney added a commit to daenney/gotosocial that referenced this pull request Jan 30, 2023
This switches the logger to Debug. By default the library logs things
internally despite returning errors. Due to the way we had it hooked up
this resulted in scary error level logs being emitted for innocuous
errors. Now those errors will only be visible in debug mode.

Fixes: superseriousbusiness#1398
Relates to: superseriousbusiness#1336
@daenney daenney mentioned this pull request Jan 30, 2023
9 tasks
tsmethurst pushed a commit that referenced this pull request Jan 31, 2023
This switches the logger to Debug. By default the library logs things
internally despite returning errors. Due to the way we had it hooked up
this resulted in scary error level logs being emitted for innocuous
errors. Now those errors will only be visible in debug mode.

Fixes: #1398
Relates to: #1336
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants