Kubernetes Autoscaling

Comprehensive autoscaling using HPA, VPA, and KEDA with kubectl-mcp-server tools.

Requires the cost, keda toolsets: start kubectl-mcp-server with --toolsets default,cost,keda (or MCP_TOOLSETS=default,cost,keda).

When to Apply

Use this skill when:

• User mentions: "HPA", "VPA", "KEDA", "autoscale", "scale to zero"
• Operations: configuring autoscaling, checking scaling status
• Keywords: "scale automatically", "event-driven", "right-size"

Priority Rules

Priority	Rule	Impact	Tools
1	Verify metrics-server for HPA	CRITICAL	kubectl_top
2	Set resource requests before HPA	CRITICAL	kubectl_describe
3	Use KEDA for scale-to-zero	HIGH	keda_scaledobjects_list
4	Check VPA recommendations	MEDIUM	get_resource_recommendations

Quick Reference

Task	Tool	Example
List KEDA ScaledObjects	keda_scaledobjects_list	keda_scaledobjects_list(namespace)
Get ScaledObject	keda_scaledobject_get	keda_scaledobject_get(name, namespace)
List ScaledJobs	keda_scaledjobs_list	keda_scaledjobs_list(namespace)
Check KEDA	detect_ecosystem	detect_ecosystem()

HPA (Horizontal Pod Autoscaler)

Basic CPU-based scaling:

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: my-app-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: my-app
  minReplicas: 2
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 70

Apply and verify:

kubectl_apply(hpa_yaml, namespace)
kubectl_get(resource_type="hpa", namespace)

VPA (Vertical Pod Autoscaler)

Right-size resource requests:

apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: my-app-vpa
spec:
  targetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: my-app
  updatePolicy:
    updateMode: "Auto"

KEDA (Event-Driven Autoscaling)

Detect KEDA Installation

detect_ecosystem()

List ScaledObjects

keda_scaledobjects_list(namespace)
keda_scaledobject_get(name, namespace)

List ScaledJobs

keda_scaledjobs_list(namespace)

Trigger Authentication

keda_triggerauths_list(namespace)
keda_triggerauth_get(name, namespace)

KEDA-Managed HPAs

keda_hpa_list(namespace)

See KEDA-TRIGGERS.md for trigger configurations.

Common KEDA Triggers

Queue-Based Scaling (AWS SQS)

apiVersion: keda.sh/v1alpha1
kind: ScaledObject
metadata:
  name: sqs-scaler
spec:
  scaleTargetRef:
    name: queue-processor
  minReplicaCount: 0
  maxReplicaCount: 100
  triggers:
  - type: aws-sqs-queue
    metadata:
      queueURL: https://sqs.region.amazonaws.com/...
      queueLength: "5"

Cron-Based Scaling

triggers:
- type: cron
  metadata:
    timezone: America/New_York
    start: 0 8 * * 1-5
    end: 0 18 * * 1-5
    desiredReplicas: "10"

Prometheus Metrics

triggers:
- type: prometheus
  metadata:
    serverAddress: http://prometheus:9090
    metricName: http_requests_total
    query: sum(rate(http_requests_total{app="myapp"}[2m]))
    threshold: "100"

Scaling Strategies

Strategy	Tool	Use Case
CPU/Memory	HPA	Steady traffic patterns
Custom metrics	HPA v2	Business metrics
Event-driven	KEDA	Queue processing, cron
Vertical	VPA	Right-size requests
Scale to zero	KEDA	Cost savings, idle workloads

Cost-Optimized Autoscaling

Scale to Zero with KEDA

Reduce costs for idle workloads:

keda_scaledobjects_list(namespace)

Right-Size with VPA

Get recommendations and apply:

get_resource_recommendations(namespace)

Troubleshooting

HPA Not Scaling

kubectl_get(resource_type="hpa", namespace)
kubectl_top(name, namespace)
kubectl_describe(name, namespace)

KEDA Not Triggering

keda_scaledobject_get(name, namespace)
kubectl_get(resource_type="events", namespace)

Common Issues

Symptom	Check	Resolution
HPA unknown	Metrics server	Install metrics-server
KEDA no scale	Trigger auth	Check TriggerAuthentication
VPA not updating	Update mode	Set updateMode: Auto
Scale down slow	Stabilization	Adjust stabilizationWindowSeconds

Best Practices

1. Always Set Resource Requests - HPA requires requests to calculate utilization
2. Use Multiple Metrics - Combine CPU + custom metrics for accuracy
3. Stabilization Windows - Prevent flapping with scaleDown stabilization
4. Scale to Zero Carefully - Consider cold start time

Related Skills

• k8s-cost - Cost optimization
• k8s-troubleshoot - Debug scaling issues