Add docs for in-place updates #7999
Conversation
This reverts commit b66b446.
…prover Add adrianmoisey to VPA approvers
…ediately after cutting a release branch so that new development is done against the new version
Bump VPA version in main branch and change release process
this change ensures that when DecreaseTargetSize is counting the nodes that it does not include any instances which are considered to be pending (i.e. not having a node ref), deleting, or are failed. this change will allow the core autoscaler to then decrease the size of the node group accordingly, instead of raising an error. This change also add some code to the unit tests to make detection of this condition easier.
…ze-fix make DecreaseTargetSize more accurate for clusterapi
Add missing tests
This change makes it so that when a failed machine is found during the `findScalableResourceProviderIDs` it will always gain a normalized provider ID with failure guard prepended. This is to ensure that machines which have gained a provider ID from the infrastructure and then later go into a failed state can be properly removed by the autoscaler when it wants to correct the size of a node group.
…-detection improve failed machine detection in clusterapi
Signed-off-by: Jack Francis <[email protected]>
Signed-off-by: Jack Francis <[email protected]>
…odeHasValidProviderID capi: node and provider ID accounting funcs
* Update default value for scaleDownDelayAfterDelete Setting defaut value for scaleDownDelayAfterDelete to be scanInterval instead of 0. * Revert the change and fix the flag description
…up-sample-scheduled Allow using scheduled pods as samples in proactive scale up
Fix log for node filtering in static autoscaler
k8s-ci-robot
added
cncf-cla: yes
|
/assign @maxcao13 Please double-check that I didn’t miss anything? I kept the docs high-level, avoiding technical details since they’re not relevant to the end user. |
Signed-off-by: Omer Aplatony <[email protected]>
|
Thanks for making this! |
maxcao13
left a comment
maxcao13
left a comment
There was a problem hiding this comment.
Looks great, have a couple suggestions. Thanks for writing this!
I think we can do that or we can just merge it afterwards, both seem reasonable to me. Maybe we should merge this after the |
Agree. |
Signed-off-by: Omer Aplatony <[email protected]>
Signed-off-by: Omer Aplatony <[email protected]>
Co-authored-by: Adrian Moisey <[email protected]>
Co-authored-by: Adrian Moisey <[email protected]>
Co-authored-by: Adrian Moisey <[email protected]>
| * All containers in a pod are updated together (partial updates not supported) | ||
| * Memory downscaling requires careful consideration to prevent OOMs | ||
| * Updates still respect VPA's standard update conditions and timing restrictions | ||
| * In-place updates will fail for pods with Guaranteed QoS class (requires pod recreation) |
There was a problem hiding this comment.
Is this true?
If the QoS class is guaranteed then requests == limits and VPA will just update both together (since the ratio between them is 1.0), which means the QoS class will never change.
My understanding is that the in-place feature will fail when you try to change the QoS class.
Am I missing something?
There was a problem hiding this comment.
You are right, my misunderstanding, sorry.
fixed in: 125209d
Signed-off-by: Omer Aplatony <[email protected]>
k8s-ci-robot
added
size/XXL
|
Gonna push this to in-place |
|
@omerap12: Closed this PR. DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
18 participants
What type of PR is this?
/kind documentation
What this PR does / why we need it:
Add docs for in place
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer:
Does this PR introduce a user-facing change?
Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.:
Depends on #7673
/hold