KREP-007: Horizontal scaling in KRO (Scale subresource + RGD Set) #866
Conversation
- Integrate KRO with the Kubernetes autoscaling ecosystem by exposing the Scale subresource on RGDs, enabling `kubectl scale`, HPA, and external autoscalers like keda. - Treat each RGD instance as a “unit of deployment”; rely on graph resolution for intra-unit ordering and propagation without progressive rollout inside a single instance. - Introduce an optional Set abstraction to replicate RGD instances and control rollout across replicas (ordering, concurrency, readiness), reducing blast radius during updates. - Simplify user workflows with familiar tooling and declarative controls while improving operator observability and separation of concerns. - Clearly scopes initial functionality (replica-count scaling with randomized names) an outlines future paths without coupling rollout logic into the core RGD reconciler.
k8s-ci-robot
added
the
cncf-cla: yes
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: tjamet The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
Hi @tjamet. Thanks for your PR. I'm waiting for a github.com member to verify that this patch is reasonable to test. If it is, they should reply with Once the patch is verified, the new status will be reflected by the I understand the commands that are listed here. 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. |
k8s-ci-robot
added
the
needs-ok-to-test
k8s-ci-robot
added
the
size/L
| - `| scale` on integer fields to represent desired/current replicas. | ||
| - `| scaleSelector` on a status string field that holds a label selector for Pods (used by HPA/VPA). |
There was a problem hiding this comment.
Firstly, this KREP is really cool.
I wonder a bit about the proliferation of keywords. WDYT about building this concept to be a bit more aware of kubernetes subresources? e.g.
integer | subresource:autoscaling/v1.Scale:specReplicasPath
integer | subresource:autoscaling/v1.Scale:statusReplicasPath
map[string]string | subresource:autoscaling/v1.Scale:labelSelectorPath
| desiredReplicas: "integer | scale" | ||
| status: | ||
| replicas: "integer | scale" | ||
| podSelector: "string | scaleSelector" |
There was a problem hiding this comment.
Noticing that a podSelector is not a string, it's a map[string]string. How would you model this?
3 participants
kubectl scale, HPA, and external autoscalers like keda.