Proposal for karpenter intergation

Author: carlory

docs/proposals/106-spot-instance-karpenter/README.md

@@ -0,0 +1,290 @@
+# Proposal-106: Support scaling with Spot instances for cost saving with Karpenter
+
+<!--
+This is the title of your Proposal. Keep it short, simple, and descriptive. A good
+title can help communicate what the Proposal is and should be considered as part of
+any review.
+-->
+
+<!--
+A table of contents is helpful for quickly jumping to sections of a Proposal and for
+highlighting any additional information provided beyond the standard Proposal
+template.
+
+Ensure the TOC is wrapped with
+  <code>&lt;!-- toc --&rt;&lt;!-- /toc --&rt;</code>
+tags, and then generate with `hack/update-toc.sh`.
+-->
+
+<!-- toc -->
+- [Proposal-106: Support scaling with Spot instances for cost saving with Karpenter](#proposal-106-support-scaling-with-spot-instances-for-cost-saving-with-karpenter)
+  - [Summary](#summary)
+  - [Motivation](#motivation)
+    - [Goals](#goals)
+    - [Non-Goals](#non-goals)
+  - [Proposal](#proposal)
+    - [User Stories (Optional)](#user-stories-optional)
+      - [Story 1: ML Engineer – Cost-Efficient Deployment of LLMs](#story-1-ml-engineer--cost-efficient-deployment-of-llms)
+      - [Story 2: Workload Author – Flexible and Preferred GPU Scheduling with Spot Support](#story-2-workload-author--flexible-and-preferred-gpu-scheduling-with-spot-support)
+    - [Notes/Constraints/Caveats (Optional)](#notesconstraintscaveats-optional)
+    - [Risks and Mitigations](#risks-and-mitigations)
+  - [Design Details](#design-details)
+    - [Test Plan](#test-plan)
+        - [Prerequisite testing updates](#prerequisite-testing-updates)
+        - [Unit tests](#unit-tests)
+        - [Integration tests](#integration-tests)
+        - [e2e tests](#e2e-tests)
+    - [Graduation Criteria](#graduation-criteria)
+  - [Implementation History](#implementation-history)
+  - [Drawbacks](#drawbacks)
+  - [Alternatives](#alternatives)
+<!-- /toc -->
+
+## Summary
+
+<!--
+This section is incredibly important for producing high-quality, user-focused
+documentation such as release notes or a development roadmap. It should be
+possible to collect this information before implementation begins, in order to
+avoid requiring implementors to split their attention between writing release
+notes and implementing the feature itself. Proposal editors and SIG Docs
+should help to ensure that the tone and content of the `Summary` section is
+useful for a wide audience.
+
+A good summary is probably at least a paragraph in length.
+
+Both in this section and below, follow the guidelines of the [documentation
+style guide]. In particular, wrap lines to a reasonable length, to make it
+easier for reviewers to cite specific portions, and to minimize diff churn on
+updates.
+
+-->
+
+
+## Motivation
+
+<!--
+This section is for explicitly listing the motivation, goals, and non-goals of
+this Proposal.  Describe why the change is important and the benefits to users. The
+motivation section can optionally provide links to [experience reports] to
+demonstrate the interest in a Proposal within the wider InftyAI community.
+
+[experience reports]: https://github.com/golang/go/wiki/ExperienceReports
+-->
+
+
+### Goals
+
+<!--
+List the specific goals of the Proposal. What is it trying to achieve? How will we
+know that this has succeeded?
+-->
+
+- Provision spot instances for inference workloads based on the model's flavor requirements.
+- Support flexible and preferred GPU scheduling with spot instances.
+- This proposal is only for AWS but the implementation can be extended to other cloud providers.
+
+### Non-Goals
+
+<!--
+What is out of scope for this Proposal? Listing non-goals helps to focus discussion
+and make progress.
+-->
+
+- Integration with the [Kubernetes Cluster Autoscaler](https://github.com/kubernetes/autoscaler) is out of scope for this proposal.
+- Add custom scheduler support for the upstream of the Karpenter project, and it is tracked in [this issue](https://github.com/kubernetes-sigs/karpenter/issues/742). Once the support is added in the upstream, we don't need to maintain the [forked version](https://github.com/InftyAI/karpenter).
+
+## Proposal
+
+<!--
+This is where we get down to the specifics of what the proposal actually is.
+This should have enough detail that reviewers can understand exactly what
+you're proposing, but should not include things like API designs or
+implementation. What is the desired outcome and how do we measure success?.
+The "Design Details" section below is for the real
+nitty-gritty.
+-->
+
+
+### User Stories (Optional)
+
+<!--
+Detail the things that people will be able to do if this Proposal is implemented.
+Include as much detail as possible so that people can understand the "how" of
+the system. The goal here is to make this feel real for users without getting
+bogged down.
+-->
+
+#### Story 1: ML Engineer – Cost-Efficient Deployment of LLMs
+
+As a machine learning engineer deploying large language models (LLMs), I don't own any physical GPU servers, so I have to rent them from cloud providers. I want to automatically use cheaper GPU Spot instances for serving models when available, so that I can reduce infrastructure costs without sacrificing performance.
+
+#### Story 2: Workload Author – Flexible and Preferred GPU Scheduling with Spot Support
+
+As a workload author, I want to publish Kubernetes manifests for my model-serving workloads that are broadly compatible across different device types, without being overly prescriptive or requiring end users to modify them.
+
+- My workloads are GPU-dependent, but there are many different GPU models available in the cloud (e.g., A10, A100, H100).
+- Instead of locking my manifests to a single GPU type, I want to express a preference-ordered list of compatible GPU types (e.g., prefer A100, fall back to A10 or L4).
+- This gives end users the flexibility to run the same manifest on different underlying infrastructure.
+- If none of the existing nodes in the cluster meet the constraints (e.g., no compatible GPUs available), I want the system to automatically provision an appropriate Spot instance from the cloud provider, based on my declared GPU preferences and resource requirements.
+
+This approach allows me to build and share portable workloads that are cost-aware, device-flexible, and production-safe, without the need for users to rewrite manifests or manage instance-level complexity themselves.
+
+### Notes/Constraints/Caveats (Optional)
+
+<!--
+What are the caveats to the proposal?
+What are some important details that didn't come across above?
+Go in to as much detail as necessary here.
+This might be a good place to talk about core concepts and how they relate.
+-->
+
+### Risks and Mitigations
+
+<!--
+What are the risks of this proposal, and how do we mitigate? Think broadly.
+For example, consider both security and how this will impact the larger
+InftyAI ecosystem.
+
+How will security be reviewed, and by whom?
+
+How will UX be reviewed, and by whom?
+
+Consider including folks who also work outside the SIG or subproject.
+-->
+
+## Design Details
+
+<!--
+This section should contain enough information that the specifics of your
+change are understandable. This may include API specs (though not always
+required) or even code snippets. If there's any ambiguity about HOW your
+proposal will be implemented, this is the place to discuss them.
+-->
+
+
+### Test Plan
+
+<!--
+**Note:** *Not required until targeted at a release.*
+The goal is to ensure that we don't accept enhancements with inadequate testing.
+
+All code is expected to have adequate tests (eventually with coverage
+expectations).
+
+[testing-guidelines]: https://git.k8s.io/community/contributors/devel/sig-testing/testing.md
+-->
+
+[x] I/we understand the owners of the involved components may require updates to
+existing tests to make this code solid enough prior to committing the changes necessary
+to implement this enhancement.
+
+##### Prerequisite testing updates
+
+<!--
+Based on reviewers feedback describe what additional tests need to be added prior
+implementing this enhancement to ensure the enhancements have also solid foundations.
+-->
+
+##### Unit tests
+
+<!--
+In principle every added code should have complete unit test coverage, so providing
+the exact set of tests will not bring additional value.
+However, if complete unit test coverage is not possible, explain the reason of it
+together with explanation why this is acceptable.
+-->
+
+<!--
+Additionally, for Alpha try to enumerate the core package you will be touching
+to implement this enhancement and provide the current unit coverage for those
+in the form of:
+- <package>: <date> - <current test coverage>
+
+This can inform certain test coverage improvements that we want to do before
+extending the production code to implement this enhancement.
+-->
+
+Forked karpenter:
+
+- `pkg/controllers/provisioning`: `Model Inference Requirements` is used to check if the model inference requirements are met when provisioning a node. And it will be added to the existing suite tests.
+
+##### Integration tests
+
+<!--
+Integration tests allow control of the configuration parameters used to start the binaries under test.
+This is different from e2e tests which do not allow configuration of parameters.
+Doing this allows testing non-default options and multiple different and potentially conflicting command line options.
+-->
+
+<!--
+This question should be filled when targeting a release.
+For Alpha, describe what tests will be added to ensure proper quality of the enhancement.
+
+For Beta and GA, add links to added tests together with links to k8s-triage for those tests:
+https://storage.googleapis.com/k8s-triage/index.html
+-->
+
+N/A.
+
+##### e2e tests
+
+<!--
+This question should be filled when targeting a release.
+For Alpha, describe what tests will be added to ensure proper quality of the enhancement.
+
+For Beta and GA, add links to added tests together with links to k8s-triage for those tests:
+https://storage.googleapis.com/k8s-triage/index.html
+
+We expect no non-infra related flakes in the last month as a GA graduation criteria.
+-->
+
+- Add one e2e test to make sure the whole system can be launched via helm chart. By leveraging kwok provider from the karpenter repo, we can test the whole system with spot instances without real cloud resources.
+- Manually test on EKS with real spot instances using custom image which is built from the forked karpenter.
+
+### Graduation Criteria
+
+<!--
+
+Clearly define what it means for the feature to be implemented and
+considered stable.
+
+If the feature you are introducing has high complexity, consider adding graduation
+milestones with these graduation criteria:
+- [Maturity levels (`alpha`, `beta`, `stable`)][maturity-levels]
+- [Feature gate][feature gate] lifecycle
+- [Deprecation policy][deprecation-policy]
+
+[feature gate]: https://git.k8s.io/community/contributors/devel/sig-architecture/feature-gates.md
+[maturity-levels]: https://git.k8s.io/community/contributors/devel/sig-architecture/api_changes.md#alpha-beta-and-stable-versions
+[deprecation-policy]: https://kubernetes.io/docs/reference/using-api/deprecation-policy/
+-->
+
+## Implementation History
+
+<!--
+Major milestones in the lifecycle of a Proposal should be tracked in this section.
+Major milestones might include:
+- the `Summary` and `Motivation` sections being merged, signaling SIG acceptance
+- the `Proposal` section being merged, signaling agreement on a proposed design
+- the date implementation started
+- the first llmaz release where an initial version of the Proposal was available
+- the version of llmaz where the Proposal graduated to general availability
+- when the Proposal was retired or superseded
+-->
+
+- 2025-06-04: Proposal drafted.
+
+## Drawbacks
+
+<!--
+Why should this Proposal _not_ be implemented?
+-->
+
+## Alternatives
+
+<!--
+What other approaches did you consider, and why did you rule them out? These do
+not need to be as detailed as the proposal, but should include enough
+information to express the idea and why it was not acceptable.
+-->

docs/proposals/106-spot-instance-karpenter/proposal.yaml

@@ -0,0 +1,35 @@
+title: Support scaling with Spot instances for cost saving
+proposal-number: 106
+authors:
+  - carlory
+status: provisional
+creation-date: 2025-06-04
+reviewers:
+  - kerthcet
+approvers:
+  - kerthcet
+
+see-also: []
+replaces: []
+
+# The target maturity stage in the current dev cycle for this proposal.
+stage: alpha
+
+# The most recent milestone for which work toward delivery of this proposal has been
+# done. This can be the current (upcoming) milestone, if it is being actively
+# worked on.
+latest-milestone: "v0.2"
+
+# The milestone at which this feature was, or is targeted to be, at each stage.
+milestone:
+  alpha: "v0.2"
+  beta: TBD
+  stable: TBD
+
+# The following PRR answers are required at alpha release
+# List the feature gate name and the components for which it must be enabled
+feature-gates: []
+disable-supported: true
+
+# The following PRR answers are required at beta release
+metrics: []