### Install and Apply Kwok Provider

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/kwok/README.md

Commands to install the Kwok provider and apply its resources to the cluster. Rerun 'make apply' if code changes.

```bash
make install-kwok
make apply # Run this command again to redeploy if the code has changed
```

--------------------------------

### Initialize E2E Integration Tests

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/dra-kwok-driver/README.md

Setup commands for the DRA integration test environment.

```bash
make setup-kind-dra     # Creates cluster, installs KWOK, deploys Karpenter
make e2etest-dra        # Runs full test suite (does all steps below automatically)
```

--------------------------------

### Install Karpenter core dependency

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/test/README.md

Add the Karpenter core repository to the Go module.

```bash
go get sigs.k8s.io/karpenter
```

--------------------------------

### Spare Capacity Delete Consolidation Example

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/balanced-consolidation.md

Example of a spare capacity delete consolidation that is approved because it saves the full node cost.

```text
savings_fraction = 4.84 / 58.08 = 8.3%
disruption_fraction = 4 / 80 = 5.0%
score = 0.083 / 0.05 = 1.67 > 0.5 --> approved
```

--------------------------------

### NodeOverlay CRD Examples

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/node-overlay.md

These examples demonstrate the structure of the NodeOverlay CRD for defining price modifications and extended resource capacity. Use these to specify targeted overrides that can be applied to NodePools.

```yaml
kind: NodeOverlay
metadata:
  name: price-modification
spec:
  requirements:
  - key: node.kubernetes.io/instance-type
    operator: In
    values: ["m5.large"]
   priceAdjustment: "+0.10"
---
kind: NodeOverlay
metadata:
  name: extended-resource
spec:
  requirements:
  - key: node.kubernetes.io/instance-type
    operator: In
    values: ["c5.large"]
  capacity:
    smarter-devices/fuse: 1
```

--------------------------------

### NodeOverlay Conflict Example

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/node-overlay.md

Example of two NodeOverlay resources with conflicting configurations and identical weights.

```yaml
kind: NodeOverlay
metadata:
  name: memory-1
spec:
  weight: 90
  requirements:
  - key: karpenter.k8s.aws/instance-memory
    operator: Gt
    values: ["m5.large", "m5.2xlarge"]
  capacity:
    memory: 50Mi
status:
  condition:
  - lastTransitionTime: "..."
    message: "conflict with overlay capacity-1"
    reason: "conflict"
    status: "False"
    type: Ready
---
kind: NodeOverlay
metadata:
  name: capacity-1
spec:
  weight: 90
  requirements:
  - key: node.kubernetes.io/instance-type
    operator: In
    values: ["c5.large", "c5.2xlarge", "m5.2xlarge"]
  capacity:
    memory: 100Mi
status:
  condition:
  - lastTransitionTime: "..."
    message: ""
    reason: ...
    status: "True"
    type: Ready
```

--------------------------------

### Well-Packed Node Consolidation Example

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/balanced-consolidation.md

Example of a well-packed node consolidation that is rejected because there are no savings.

```text
One m7i.xlarge runs 10 pods requesting 3.5 vCPU and 14 GiB. The smallest fitting replacement is another m7i.xlarge. Savings is $0. No threshold approves this move.
```

--------------------------------

### Node Status Example: Ready Condition False

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/node-repair.md

An example Kubernetes Node object showing both 'NetworkUnavailable' and 'Ready' conditions with a status of 'False', indicating potential node unreadiness.

```yaml
---
apiVersion: v1
kind: Node
metadata:
  ...
status:
  condition:
    - lastHeartbeatTime: "2024-11-01T16:29:49Z"
      lastTransitionTime: "2024-11-01T15:02:48Z"
      message: kubelet is posting ready status
      reason: KubeletReady
      status: "False"
      type: NetworkUnavailable
    - lastHeartbeatTime: "2024-11-01T16:29:49Z"
      lastTransitionTime: "2024-11-01T15:02:48Z"
      message: kubelet is posting ready status
      reason: KubeletReady
      status: "False"
      type: Ready
    ...
    
- The Node here will be eligible for node repair after at `2024-11-01T15:32:48Z`
```

--------------------------------

### Define Flexibility Requirements

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/nodepool-requirement-flexibility.md

Example configuration demonstrating conflicting flexibility requirements for instance generation and category.

```yaml
requirements:
- key: karpenter.kwok.sh/instance-generation
  operator: Exists
  flexibility: 1
- key: karpenter.kwok.sh/instance-category
  operator: Exists
  flexibility: 2
```

--------------------------------

### Oversized Node Consolidation Example

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/balanced-consolidation.md

Example demonstrating an oversized node consolidation that is approved due to significant savings.

```text
savings_fraction = 7.26 / 58.08 = 12.5%
disruption_fraction = 3 / 80 = 3.75%
score = 0.125 / 0.0375 = 3.33 > 0.5 --> approved
```

--------------------------------

### Define NodeOverlay resources

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/node-overlay.md

Example showing multiple NodeOverlay configurations with varying weights and requirements for capacity and pricing adjustments.

```yaml
kind: NodeOverlay
metadata:
  name: default
spec:
  priceAdjustment: "-10%"
  capacity:
    memory: 10Mi
---
kind: NodeOverlay
metadata:
  name: memory
spec:
  weight: 1 # required to resolve conflicts with default
  requirements:
  - key: karpenter.k8s.aws/instance-memory
    operator: Gt
    values: [2048]
  capacity:
    memory: 50Mi
---
kind: NodeOverlay
metadata:
  name: extended-resource
spec:
  # weight is not required since default does not conflict with extended-resource
  requirements:
  - key: node.kubernetes.io/instance-type
    operator: In
    values: ["m5.large", "m5.2xlarge", "m5.4xlarge", "m5.8xlarge", "m5.12xlarge"]
  capacity:
    smarter-devices/fuse: 1
```

--------------------------------

### Configure GitHub Actions workflow

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/test/README.md

Example workflow configuration for running integration tests on push or pull requests.

```yaml
# .github/workflows/tests.yaml
name: Integration Tests
on:
  push:
    branches: [ main ]
  pull_request:
    branches: [ main ]

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - name: Set up Go
        uses: actions/setup-go@v4
        with:
          go-version: '1.20'
      - name: Set up cloud provider credentials
        run: |
          # Set up your cloud provider credentials here
      - name: Run tests
        run: make upstream-e2etests
```

--------------------------------

### Marginal Move Consolidation Example

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/balanced-consolidation.md

Example of a marginal consolidation move that is rejected due to low savings relative to disruption.

```text
savings_fraction = 2.42 / 58.08 = 4.2%
disruption_fraction = 8 / 80 = 10.0%
score = 0.042 / 0.10 = 0.42 < 0.5 --> rejected
```

--------------------------------

### NodePool Disruption Spec Example

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/disruption-controls.md

Example of the Disruption block within a v1beta1 NodePool spec, demonstrating consolidation policy, consolidation after duration, expiration, and budget configurations.

```yaml
apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
  name: default
spec: # This is not a complete NodePool Spec.
  disruption:
    consolidationPolicy: WhenUnderutilized || WhenEmpty
    consolidateAfter: 10m || Never # metav1.Duration
    expireAfter: 10m || Never # Equivalent to v1alpha5 TTLSecondsUntilExpired
    budgets:
    # On Weekdays during business hours, don't do any deprovisioning.
    - schedule: "0 9 * * mon-fri"
      duration: 8h
      nodes: 0
    # Every other time, only allow 10 nodes to be deprovisioned simultaneously
    - nodes: 10
```

--------------------------------

### Configure NodePool Disruption and Budget

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/disruption-controls.md

Example of a NodePool specification with default consolidation policy and a 10% node budget.

```yaml
apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
  name: default
spec:
  disruption:
    consolidationPolicy: WhenUnderutilized
    consolidateAfter: 15s
    expireAfter: 30d
    budgets:
    - nodes: 10%
```

--------------------------------

### NodePool with minValues for Instance Family and Type

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/nodepool-requirement-flexibility.md

This example demonstrates how to configure a NodePool with `minValues` for instance family and instance type requirements. It ensures that at least 2 instance families and 10 instance types are available when scheduling pods.

```yaml
apiVersion: karpenter.sh/v1beta1
kind: NodePool
spec:
  template:
    spec:
      requirements:
        - key: karpenter.kwok.sh/instance-family
          operator: In
          values: ["c", "m", "r"]
          minValues: 2
        - key: node.kubernetes.io/instance-type
          operator: Exists
          minValues: 10
```

--------------------------------

### Configure Multi-Reason Disruption Budgets

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/disruption-controls-by-reason.md

Example of a NodePool specification using multiple reasons within a single budget entry.

```yaml
apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
  name: default
spec: # This is not a complete NodePool Spec.
  disruption:
    budgets:
    - schedule: "* * * * *"
      reasons: [Drifted, Underutilized]
      nodes: 10
    # For all other reasons, only allow 5 nodes to be disrupted at a time
    - nodes: 5
      schedule: "* * * * *"
```

--------------------------------

### Define Per Reason Controls in NodePool

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/disruption-controls-by-reason.md

Example configuration for a NodePool using per-reason disruption controls, including budgets for consolidation and drift.

```yaml
apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
  name: example-nodepool
spec:
  disruption:
    defaults:
      budgets: 
        - nodes: 10% 
          schedule: "0 0 1 * *"
          duration: 1h 
    consolidation:
      consolidationPolicy: WhenUnderutilized
      disruptAfter: "30m"
    drift:
      budgets:
        - nodes: "20%"
          schedule: "0 0 1 * *"
          duration: "1h"
        - nodes: "10%"
          schedule: "0 0 * * 0"
          duration: "2h"
        - nodes: "50%" 
          schedule: "@monthly"
```

--------------------------------

### Configure Single-Reason Disruption Budgets

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/disruption-controls-by-reason.md

Example of a NodePool specification where each budget entry is limited to a single reason.

```yaml
apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
  name: default
spec: # This is not a complete NodePool Spec.
  disruption:
    budgets:
    - schedule: "* * * * *"
      reason: drifted 
      nodes: 10
    # For all other reasons , only allow 5 nodes to be disrupted at a time
    - nodes: 5
      schedule: "* * * * *"
```

--------------------------------

### Configure Multi-Driver DRA Setup

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/dra-kwok-driver/README.md

Simulates multiple DRA drivers (GPU and FPGA) on the same nodes using separate DRAConfig resources.

```yaml
# GPU driver simulation
apiVersion: test.karpenter.sh/v1alpha1
kind: DRAConfig
metadata:
  name: gpu-config  # User-chosen name
  namespace: karpenter
spec:
  driver: gpu.nvidia.com  # Simulated driver name
  pools:
    - name: h100-pool
      nodeSelectorTerms:
        - matchExpressions:
            - key: node.kubernetes.io/instance-type
              operator: In
              values: ["g5.xlarge"]
      resourceSlices:
        - devices:
            - name: h100-0
              attributes:
                type: {stringValue: "nvidia-h100"}
                memory: {stringValue: "80Gi"}

---
# FPGA driver simulation
apiVersion: test.karpenter.sh/v1alpha1
kind: DRAConfig
metadata:
  name: fpga-config
  namespace: karpenter
spec:
  driver: fpga.intel.com  # Different driver
  pools:
    - name: arria-pool
      nodeSelectorTerms:
        - matchExpressions:
            - key: node.kubernetes.io/instance-type
              operator: In
              values: ["g5.xlarge"]  # Same instance type as above
      resourceSlices:
        - devices:
            - name: arria-0
              attributes:
                type: {stringValue: "intel-arria10"}
```

--------------------------------

### Example Node Overlay with DRA Configuration

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/karpenter-dra-kwok-driver.md

This YAML defines a NodeOverlay configuration for testing Karpenter's integrated DRA scheduling. It specifies requirements for instance type and custom extended resources for DRA devices, with future support for embedding ResourceSlice objects.

```yaml
apiVersion: test.karpenter.sh/v1alpha1
kind: NodeOverlay
metadata:
  name: gpu-dra-config
spec:
  weight: 10  # Higher weight for conflict resolution
  requirements:
  - key: node.kubernetes.io/instance-type
    operator: In
    values: ["g5.48xlarge"]
  capacity:
    test.karpenter.sh/device: "8"  # Custom extended resource for DRA devices
  # TODO: Extend NodeOverlay API to embed ResourceSlice templates
  resourceSlices:  # FUTURE: Embedded ResourceSlice objects (not yet implemented)
  - apiVersion: resource.k8s.io/v1
    kind: ResourceSlice
    spec:
      # nodeName will be filled in by driver when node is created
      driver: "test.karpenter.sh"
      devices:
      - name: "nvidia-h100-0"
        driver: "test.karpenter.sh"
        attributes:
          memory: "80Gi"
          compute-capability: "9.0"
          vendor: "nvidia"
      - name: "nvidia-h100-1"
        driver: "test.karpenter.sh"
        attributes:
          memory: "80Gi"
          compute-capability: "9.0"
          vendor: "nvidia"

```

--------------------------------

### Manage DRA Integration Test Lifecycle

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/dra-kwok-driver/README.md

Individual commands used during the E2E integration test process for setup, execution, and cleanup.

```bash
kubectl apply -f dra-kwok-driver/pkg/apis/crds/test.karpenter.sh_draconfigs.yaml
```

```bash
kubectl delete draconfigs --all -n karpenter
kubectl delete resourceslices --all
pkill -f dra-kwok-driver  # Kills running driver binary
```

```bash
cd dra-kwok-driver && go build -o dra-kwok-driver main.go
```

```bash
./dra-kwok-driver > /tmp/dra-driver.log 2>&1 &
```

```bash
TEST_SUITE=dra make e2etests
```

```bash
kubectl delete draconfigs --all
kubectl delete resourceslices --all
kill <dra-driver-pid>
```

--------------------------------

### NodePool with Balanced Consolidation Policy

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/balanced-consolidation.md

Example of a NodePool configuration specifying balanced consolidation. This policy enables consolidation that considers within-family replacements.

```yaml
apiVersion: karpenter.sh/v1
kind: NodePool
metadata:
  name: default
spec:
  disruption:
    consolidationPolicy: Balanced
    consolidateAfter: 30s
    budgets:
    - nodes: 10%
```

--------------------------------

### NodePool with 72-hour Eventual Disruption

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/do-not-disrupt-grace-period.md

This example shows a NodePool configuration with a `terminationGracePeriod` of 72 hours. When combined with pod-level `do-not-disrupt` annotations, it ensures that nodes are eventually disrupted even if pods have protection, with pod protection expiring first or node protection expiring.

```yaml
spec:
  disruption:
    consolidationPolicy: WhenEmpty
  template:
    spec:
      terminationGracePeriod: 72h
```

--------------------------------

### Initialize test directories

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/test/README.md

Create the necessary directory structure for provider-specific test configurations.

```bash
mkdir -p test/suites
mkdir -p test/pkg/environment/yourprovider
```

--------------------------------

### Node Status Example: Network Unavailable

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/node-repair.md

An example Kubernetes Node object demonstrating the 'NetworkUnavailable' condition with a status of 'False', indicating a network issue.

```yaml
apiVersion: v1
kind: Node
metadata:
  ...
status:
  condition:
    - lastHeartbeatTime: "2024-11-01T16:29:49Z"
      lastTransitionTime: "2024-11-01T15:02:48Z"
      message: no connection
      reason: Network is not available
      status: "False"
      type: NetworkUnavailable
    ...
    
- The Node here will be eligible for node repair after at `2024-11-01T15:12:48Z`
```

--------------------------------

### NodeRegistrationHealthy Status Condition Example

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/noderegistrationhealthy-status-condition.md

Representation of the NodeRegistrationHealthy condition within the NodePool status object.

```yaml
// 'NodeRegistrationHealthy' condition indicates if a misconfiguration exists that prevents the normal, successful use of a Karpenter resource
Status:
  Conditions:
    Last Transition Time:  2025-01-13T18:57:20Z
    Message:               
    Observed Generation:   1
    Reason:                NodeRegistrationHealthy
    Status:                True
    Type:                  NodeRegistrationHealthy
```

--------------------------------

### Karpenteradm Diagnose Command with Overlays

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/node-overlay.md

This command demonstrates how to use `karpenteradm diagnose` to preview the impact of Node Overlay configurations on instance pricing and capacity. It helps visualize changes before deployment.

```bash
> karpenteradm diagnose --type instance-type --overlay price-modification extended-resource --nodepool default
NODEPOOL    OVERLAY                      INSTANCE_TYPE   CAPACITY TYPE      PRICE    CAPCITY
default     "price-modification"         m5.large        spot               0.47     
default     "extended-resource"          c5.large        spot               0         smarter-devices/fuse: 1
...

```

--------------------------------

### NodeClaim API Definition

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/v1-api.md

Example of a NodeClaim resource definition including spec requirements and status conditions.

```yaml
apiVersion: karpenter.sh/v1
kind: NodeClaim
metadata:
  name: default
spec:
  nodeClassRef:
    group: karpenter.k8s.aws # Updated since only a single version will be served
    kind: EC2NodeClass
    name: default
  taints:
    - key: example.com/special-taint
      effect: NoSchedule
  startupTaints:
    - key: example.com/another-taint
      effect: NoSchedule
  requirements:
    - key: "karpenter.k8s.aws/instance-category"
      operator: In
      values: ["c", "m", "r"]
      minValues: 2
  resources:
    requests:
      cpu: "20"
      memory: "8192Mi"
  expireAfter: 720h | Never
  terminationGracePeriod: 1d
status:
  allocatable:
    cpu: 1930m
    ephemeral-storage: 17Gi
    memory: 3055Mi
    pods: "29"
    vpc.amazonaws.com/pod-eni: "9"
  capacity:
    cpu: "2"
    ephemeral-storage: 20Gi
    memory: 3729Mi
    pods: "29"
    vpc.amazonaws.com/pod-eni: "9"
  conditions:
    - lastTransitionTime: "2024-02-02T19:54:34Z"
      status: "True"
      type: Launched
    - lastTransitionTime: "2024-02-02T19:54:34Z"
      status: "True"
      type: Registered
    - lastTransitionTime: "2024-02-02T19:54:34Z"
      status: "True"
      type: Initialized
    - lastTransitionTime: "2024-02-02T19:54:34Z"
      status: "True"
      type: Drifted
      reason: RequirementsDrifted
    - lastTransitionTime: "2024-02-02T19:54:34Z"
      status: "True"
      type: Expired
    - lastTransitionTime: "2024-02-02T19:54:34Z"
      status: "True"
      type: Empty
    - lastTransitionTime: "2024-02-02T19:54:34Z"
      status: "True"
      type: Ready
  nodeName: ip-192-168-71-87.us-west-2.compute.internal
  providerID: aws:///us-west-2b/i-053c6b324e29d2275
  imageID: ami-0b1e393fbe12f411c
```

--------------------------------

### Configure Static NodePool with Scaling and Disruption Budgets

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/static-capacity.md

A complete NodePool configuration demonstrating fixed replica counts, instance type constraints, expiration, and time-based disruption budgets.

```yaml
apiVersion: karpenter.sh/v1beta1
kind: NodePool
metadata:
  name: static-prod-nodepool
spec:
  replicas: 12 # Maintains 12 nodes
  template:
    metadata:
      labels:
        nodepool-type: static
    spec:
      nodeClassRef:
        group: karpenter.k8s.aws
        kind: EC2NodeClass
        name: myEc2NodeClass
      expireAfter: 720h
      taints:
        - key: example.com/special-taint
          effect: NoSchedule
      requirements:
        - key: topology.kubernetes.io/zone
          operator: In
          values: ["zone-2a"]
        - key: karpenter.k8s.aws/instance-type
          operator: In
          values: ["m5.2xlarge"]
  disruption:
    budgets:
      - nodes: 10%
      # On Weekdays during business hours, don't do any deprovisioning.
      - schedule: "0 9 * * mon-fri"
        duration: 8h
        nodes: "0"
  limits:
    # Constrain the size of static Nodepool 
    nodes: 20 # Karpenter will not scale up nodes or create replacements if limits are breaching
```

--------------------------------

### Define Zonal Topology Requirements

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/static-capacity.md

Example of specifying multiple availability zones within a NodePool requirement.

```yaml
- key: topology.kubernetes.io/zone
  operator: In
  values: ["zone-2a", "zone-2b", "zone-2c"]
```

--------------------------------

### NodeClaim Printer Columns

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/v1-api.md

Defines the columns displayed when listing NodeClaims using `kubectl get nodeclaims`.

```APIDOC
## NodeClaim Printer Columns

### Description
Printer columns define the output format for `kubectl get nodeclaims`. They can be customized to show specific fields.

### Standard Columns
1. **Name**: The name of the NodeClaim.
2. **Instance Type**: The instance type provisioned for the NodeClaim.
3. **Capacity**: The capacity of the provisioned node (e.g., spot, on-demand).
4. **Zone**: The availability zone where the node is provisioned.
5. **Node**: The name of the Kubernetes node.
6. **Ready**: Indicates if the node is ready to accept pods.
7. **Age**: The age of the NodeClaim resource.

### Wide Columns (-o wide)
In addition to standard columns, wide output includes:
1. **ID**: The provider ID of the node (e.g., AWS instance ID).
2. **NodePool Name**: The name of the NodePool associated with the NodeClaim.
3. **NodeClass Name**: The name of the NodeClass used for provisioning.
```

--------------------------------

### Run End-to-End Tests

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/kwok/README.md

Command to execute all end-to-end tests for the Kwok provider. Set the FOCUS environment variable to test a specific case.

```bash
export FOCUS="<e2e case name>"
make e2etests
```

--------------------------------

### NodePool API Specification

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/v1-api.md

Example of a NodePool resource definition including template, disruption budgets, and status fields.

```yaml
apiVersion: karpenter.sh/v1
kind: NodePool
metadata:
  name: default
spec:
  template:
    metadata:
      labels:
        billing-team: my-team
      annotations:
        example.com/owner: "my-team"
    spec:
      nodeClassRef:
        group: karpenter.k8s.aws # Updated since only a single version will be served
        kind: EC2NodeClass
        name: default
      taints:
        - key: example.com/special-taint
          effect: NoSchedule
      startupTaints:
        - key: example.com/another-taint
          effect: NoSchedule
      requirements:
        - key: "karpenter.k8s.aws/instance-category"
          operator: In
          values: ["c", "m", "r"]
          minValues: 2 # Alpha field, added for spot best practices support
      expireAfter: 720h | Never
      terminationGracePeriod: 1d
  disruption:
    budgets:
      - nodes: 0
        schedule: "0 10 * * mon-fri"
        duration: 16h
        reasons:
          - Drifted
          - Expired
      - nodes: 100%
        reasons:
          - Empty
      - nodes: "10%"
      - nodes: 5
    consolidationPolicy: WhenUnderutilized | WhenEmpty
    consolidateAfter: 1m | Never # Added to allow additional control over consolidation aggressiveness
  weight: 10
  limits:
    cpu: "1000"
    memory: 1000Gi
status:
  conditions:
    - lastTransitionTime: "2024-02-02T19:54:34Z"
      status: "True"
      type: NodeClassReady
    - lastTransitionTime: "2024-02-02T19:54:34Z"
      status: "True"
      type: Ready
  resources:
    nodes: "5"
    cpu: "20"
    memory: "8192Mi"
    ephemeral-storage: "100Gi"
```

--------------------------------

### Active Buffer with Resource Limits

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/capacity-buffers.md

Configure a Capacity Buffer with a PodTemplate and specific resource limits for CPU and memory. This ensures that the buffer does not exceed defined resource constraints.

```yaml
apiVersion: v1
kind: PodTemplate
metadata:
  name: worker-template
  namespace: workers
template:
  spec:
	containers:
	- name: worker
	  image: worker:latest
	  resources:
		requests:
		  cpu: "2"
		  memory: "4Gi"
---
apiVersion: autoscaling.x-k8s.io/v1alpha1
kind: CapacityBuffer
metadata:
  name: worker-buffer
  namespace: workers
spec:
  provisioningStrategy: "buffer.x-k8s.io/active-capacity"
  podTemplateRef:
	name: worker-template
  limits:
	cpu: "20"
	memory: "40Gi"
```

--------------------------------

### NodePool Printer Columns

Source: https://github.com/kubernetes-sigs/karpenter/blob/main/designs/v1-api.md

Defines the columns displayed when listing NodePool resources using `kubectl get nodepools`.

```APIDOC
## NodePool Printer Columns

### Description
Columns displayed for NodePool resources in `kubectl get nodepools` output, including standard and wide views.

### Current Columns (Standard)
- **NAME**
- **NODECLASS**
- **WEIGHT**

### Proposed Columns (Standard)
- **Name**
- **NodeClass**
- **Nodes**
- **Ready**
- **Age**
- **Weight**

### Proposed Columns (Wide - `-o wide`)
- **Weight**
- **CPU**
- **Memory**
```