[PLAT-713] Allow the CFS period to be tuned for containers #1
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
guillaumebreton
approved these changes
Aug 10, 2017
pkg/kubelet/cm/helpers_linux.go
Outdated
| ) | ||
|
|
||
| // MilliCPUToQuota converts milliCPU to CFS quota and period values. | ||
| func MilliCPUToQuota(milliCPU int64) (quota int64, period int64) { | ||
| // MilliCPUToQuota takes milliCPU (along with a CFS period, in usec) and retuns |
|
Merged into release/1.7 instead. |
jackkleeman
pushed a commit
that referenced
this pull request
Feb 14, 2019
update from base repository
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Kubernetes' inbuilt resource limits allow control over the maximum CPU bandwidth a container can consume. "Under the hood," this is achieved by setting
cpu.cfs_quotaon a container's cgroup. We have been setting these limits for most of our deployments, and they are effective at preventing "greedy" containers from negatively impacting the performance of other co-located containers 👮However, when a container gets throttled, it is unable to run on the CPU again until the end of the "throttling period." (See golang/go#19378.) The length of this period is controlled by the
cpu. cfs_period_usparameter and is currently set statically to 100ms. Services on the "serving path" tend to be very latency sensitive – effectively being paused for 100ms is extremely detrimental to performance and can lead to very high tail latencies.This change makes this period configurable by way of a
monzo.com/cpu-periodparameter in a container's resource limits, specified in microseconds.