feat: support custom instance types with the KWOK provider

[drmorr0]

Fixes #987

Description

This extends the KWOK provider for Karpenter by allowing users to specify custom lists of instance types that will be launched by Karpenter/KWOK.

How was this change tested?
Manual testing in a local kind+kwok cluster; I confirmed that, with the provided kwok/examples/instance_types.json, that an nginx deployment with zonal pod topology spread enabled worked correctly:

> cat test.k8s.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: test
  name: test-depl
  namespace: simkube
spec:
  replicas: 3
  selector:
    matchLabels:
      app: test
  template:
    metadata:
      labels:
        app: test
    spec:
      containers:
        - env:
            - name: POD_OWNER
              value: test-depl
          image: nginx:latest
          name: nginx
          resources:
            requests:
              cpu: "1"
      nodeSelector:
        type: virtual
      tolerations:
        - effect: NoSchedule
          key: kwok-provider
          value: "true"
      topologySpreadConstraints:
        - maxSkew: 1
          topologyKey: "topology.kubernetes.io/zone"
          whenUnsatisfiable: DoNotSchedule
          labelSelector:
            matchLabels:
              app: test
> kubectl get nodeclaims
NAME                 TYPE         ZONE         NODE                        READY   AGE
kwok-default-9szxv   g5.4xlarge   us-east-1c   strange-swirles-833325525   True    4m41s
kwok-default-twk5m   g5.4xlarge   us-east-1b   strange-joliot-1822046751   True    4m41s
kwok-default-vrcqw   g5.4xlarge   us-east-1a   kind-ganguly-3882149666     True    4m41s
> kubectl get nodes
NAME                        STATUS   ROLES           AGE     VERSION
karpenter-control-plane     Ready    control-plane   15d     v1.29.0
karpenter-worker            Ready    <none>          15d     v1.29.0
kind-ganguly-3882149666     Ready    <none>          5m18s   kwok-v0.4.0
strange-joliot-1822046751   Ready    <none>          5m18s   kwok-v0.4.0
strange-swirles-833325525   Ready    <none>          5m18s   kwok-v0.4.0
> kubectl get pods
NAMESPACE   NAME                             READY   STATUS    RESTARTS   AGE
simkube     test-depl-c5f8ddd84-h5tzw        1/1     Running   0          5m58s
simkube     test-depl-c5f8ddd84-nnz9x        1/1     Running   0          5m58s
simkube     test-depl-c5f8ddd84-xshgx        1/1     Running   0          5m58s

Here are the relevant logs from karpenter:

{"level":"INFO","time":"2024-02-27T23:00:02.648Z","logger":"controller.provisioner","message":"computed new nodeclaim(s) to fit pod(s)","commit":"7b5bd17-dirty","nodeclaims":3,"pods":3}
{"level":"INFO","time":"2024-02-27T23:00:02.666Z","logger":"controller.provisioner","message":"created nodeclaim","commit":"7b5bd17-dirty","nodepool":"kwok-default","nodeclaim":"kwok-default-9szxv","requests":{"cpu":"1","pods":"2"},"instance-types":"g5.4xlarge"}
{"level":"INFO","time":"2024-02-27T23:00:02.667Z","logger":"controller.provisioner","message":"created nodeclaim","commit":"7b5bd17-dirty","nodepool":"kwok-default","nodeclaim":"kwok-default-vrcqw","requests":{"cpu":"1","pods":"2"},"instance-types":"g5.4xlarge"}
{"level":"INFO","time":"2024-02-27T23:00:02.667Z","logger":"controller.provisioner","message":"created nodeclaim","commit":"7b5bd17-dirty","nodepool":"kwok-default","nodeclaim":"kwok-default-twk5m","requests":{"cpu":"1","pods":"2"},"instance-types":"g5.4xlarge"}
{"level":"INFO","time":"2024-02-27T23:00:02.683Z","logger":"controller.nodeclaim.lifecycle","message":"launched nodeclaim","commit":"7b5bd17-dirty","nodeclaim":"kwok-default-vrcqw","provider-id":"kwok://kind-ganguly-3882149666","instance-type":"g5.4xlarge","zone":"us-east-1a","capacity-type":"on-demand","allocatable":{"cpu":"15900m","ephemeral-storage":"600Gi","m
emory":"65526Mi","nvidia.com/gpu":"1","pods":"110"}}
{"level":"INFO","time":"2024-02-27T23:00:02.685Z","logger":"controller.nodeclaim.lifecycle","message":"launched nodeclaim","commit":"7b5bd17-dirty","nodeclaim":"kwok-default-twk5m","provider-id":"kwok://strange-joliot-1822046751","instance-type":"g5.4xlarge","zone":"us-east-1b","capacity-type":"on-demand","allocatable":{"cpu":"15900m","ephemeral-storage":"600Gi",
"memory":"65526Mi","nvidia.com/gpu":"1","pods":"110"}}
{"level":"INFO","time":"2024-02-27T23:00:02.687Z","logger":"controller.nodeclaim.lifecycle","message":"launched nodeclaim","commit":"7b5bd17-dirty","nodeclaim":"kwok-default-9szxv","provider-id":"kwok://strange-swirles-833325525","instance-type":"g5.4xlarge","zone":"us-east-1c","capacity-type":"on-demand","allocatable":{"cpu":"15900m","ephemeral-storage":"600Gi",
"memory":"65526Mi","nvidia.com/gpu":"1","pods":"110"}}

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[coveralls]

Pull Request Test Coverage Report for Build 8976607133

Details

• 0 of 138 (0.0%) changed or added relevant lines in 4 files are covered.
• 2 unchanged lines in 2 files lost coverage.
• Overall coverage decreased (-0.3%) to 78.566%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
kwok/main.go	0	6	0.0%
kwok/cloudprovider/cloudprovider.go	0	14	0.0%
kwok/cloudprovider/helpers.go	0	52	0.0%
kwok/tools/gen_instance_types.go	0	66	0.0%

Files with Coverage Reduction	New Missed Lines	%
kwok/cloudprovider/cloudprovider.go	1	0.0%
kwok/cloudprovider/helpers.go	1	0.0%

Totals
Change from base Build 8975908762:	-0.3%
Covered Lines:	8350
Relevant Lines:	10628

---

💛 - Coveralls

[njtran]

Super nice work! Love the simplifications you've made.

[drmorr0]

@njtran I've responded to your comments and committed some changes, lmk what you think! If these look good I can rebase them into the stack.

The major update is that, as suggested, we're now just embedding the data into the binary, so I've reverted all the changes to the helm charts.

[drmorr0]

@njtran any other concerns or comments about this PR? Did my response about the spot/on-demand logic make sense?

[njtran]

/hold

[njtran]

/lgtm
/approve
Will wait for @jonathan-innis to take a look

[njtran]

/lgtm cancel

Oops got the command wrong

[njtran]

/remove-approve

[drmorr0]

Added the boilerplate to gen_instance_types.go

[drmorr0]

@jonathan-innis OK I think I've addressed the remaining comments in here? I tested this again locally and it seems to work, and tests here pass as well. If there are additional concerns, I'd kinda like to address those in follow-up PRs if that's OK? There's been quite a bit of back-and-forth on this already.

[njtran]

/lgtm
/approve

Good work! Thanks for committing this!

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: drmorr0, njtran

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [njtran]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[njtran]

/remove-hold