docs: Support Capacity Buffers in Karpenter#2898
docs: Support Capacity Buffers in Karpenter#2898sumukha-radhakrishna wants to merge 7 commits intokubernetes-sigs:mainfrom
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: sumukha-radhakrishna The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
k8s-ci-robot
added
the
cncf-cla: yes
k8s-ci-robot
added
the
needs-ok-to-test
|
Hi @sumukha-radhakrishna. Thanks for your PR. I'm waiting for a kubernetes-sigs member to verify that this patch is reasonable to test. If it is, they should reply with Tip We noticed you've done this a few times! Consider joining the org to skip this step and gain Once the patch is verified, the new status will be reflected by the I understand the commands that are listed here. DetailsInstructions 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. |
k8s-ci-robot
added
the
size/XL
sumukha-radhakrishna
changed the title
sumukha-radhakrishna
force-pushed
the
capacity-buffers-rfc
branch
from
cf21be3 to
7888e5a
Compare
|
|
||
| **Ephemeral Capacity Strategy:** | ||
|
|
||
| After the initial active buffer implementation, we will implement ephemeral capacity strategy to support batch systems like Kueue. Ephemeral strategy will provide: |
There was a problem hiding this comment.
Nit: To harden support for batch systems. Kueue can work with active buffers, its just a bit racey
| - Ephemeral capacity strategy: One-time capacity request pattern for batch systems like Kueue (deferred to future work) | ||
| - Adding `expireAfter` field to the API in initial implementation (requires upstream sig-autoscaling consensus) | ||
|
|
||
| # Future Work |
There was a problem hiding this comment.
Nit: this should probably all be towards the end, we want to talk about what we are working on first instead of what we will work on
| - No → Create NodeClaims, keep `Provisioning: False` until nodes are available | ||
| 4. Only sets buffer status to `Provisioning: True` when virtual pods can be successfully placed on existing cluster capacity without creating new NodeClaims | ||
|
|
||
| **Key Point:** Virtual pods are reconstructed every provisioning loop from buffer status. No pod objects are created. The `Provisioning: True` status reflects actual available capacity in the cluster, ensuring the status accurately represents whether buffer capacity is ready for use even if NodeClaims fail to provision. |
There was a problem hiding this comment.
I assume we'll cache the recreated pods?
4 participants
Fixes #749 #2571
Description
Support sig-autoscaling CapacityBuffer API
How was this change tested?
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.