KUBE-5984 - ensure that scale-ups always occur when there are starved pods #225
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
If you want this to close the issue, you'll need to use the keywords from here - https://docs.github.com/en/issues/tracking-your-work-with-issues/linking-a-pull-request-to-an-issue#linking-a-pull-request-to-an-issue-using-a-keyword, e.g. closes/fixes/resolves etc. |
awprice
requested changes
Jun 23, 2023
|
Would be good to see in the description of this PR what the change does to fix the issue - I had to look through the code to see how it addresses the original issue |
hittingray
reviewed
Jun 25, 2023
awprice
previously approved these changes
Jun 27, 2023
|
LGTM - thanks for the contribution! |
Jacobious52
previously approved these changes
Jun 30, 2023
Jacobious52
approved these changes
Jul 2, 2023
awprice
approved these changes
Jul 2, 2023
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Outdated suggestions cannot be applied.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.
This fixes #224
The main change here is to add a
ScaleOnStarveoption to the node group configuration. This true/false value configures an additional check on the nodeDelta calculated during the scaling step.When we gather the RequestedPod, we also gather the largest pending pods (by both CPU and Memory). When we gather the node capacity, we also gather the largest node (node with allocatable CPU/Memory minus used pod CPU/memory) that is the highest. If either of the requested pods have larger requirements than what's available on the largest capacity, then that indicates we have a "starved pod". In the case that a pod exists with no nodes available (and we have
ScaleOnStarveenabled), then we make sure that the scaling algorithm has at least 1 scale up as the final result.