Mastering Scheduled Autoscaling in Kubernetes with cronhpa-controller

Background

Container orchestration simplifies software delivery and operations, but static capacity planning often wastes resources. Typical web workloads show peak‑to‑valley traffic differences of 3‑4× or even 10×, leading to waste ratios of 25%–57%.

The standard Horizontal Pod Autoscaler (HPA) reacts to metric thresholds (CPU, memory, custom metrics) but incurs minute‑level latency in data collection, decision making, and scaling. This latency is problematic for workloads with sharp spikes, such as online games.

kubernetes‑cronhpa‑controller

The kubernetes-cronhpa-controller is an open‑source CRD‑based controller that enables time‑based scaling schedules. By defining cron‑like jobs, users can pre‑scale resources before predicted peaks and shrink them during troughs. When combined with cluster‑autoscaler, node‑level scaling further reduces infrastructure cost.

Installation

Install the CRD

kubectl apply -f config/crds/autoscaling_v1beta1_cronhorizontalpodautoscaler.yaml

Grant RBAC permissions

# Create ClusterRole
kubectl apply -f config/rbac/rbac_role.yaml
# Create ClusterRoleBinding and ServiceAccount
kubectl apply -f config/rbac/rbac_role_binding.yaml

Deploy the controller kubectl apply -f config/deploy/deploy.yaml Verify installation

kubectl get deploy kubernetes-cronhpa-controller -n kube-system -o wide

Demo

Create a CronHorizontalPodAutoscaler resource that scales a sample deployment up to 3 replicas at the start of each minute and down to 1 replica at the 30‑second mark.

apiVersion: autoscaling.alibabacloud.com/v1beta1
kind: CronHorizontalPodAutoscaler
metadata:
  name: cronhpa-sample
  namespace: default
spec:
  scaleTargetRef:
    apiVersion: apps/v1beta2
    kind: Deployment
    name: nginx-deployment-basic
  jobs:
  - name: "scale-down"
    schedule: "30 */1 * * * *"
    targetSize: 1
  - name: "scale-up"
    schedule: "0 */1 * * * *"
    targetSize: 3

Deploy the demo resources and observe the scaling behavior:

# Apply the CronHPA and the demo deployment
kubectl apply -f examples/deployment_cronhpa.yaml

# Check replica count of the demo deployment
kubectl get deploy nginx-deployment-basic

# Describe the CronHPA to see job status
kubectl describe cronhpa cronhpa-sample

The describe output shows the jobs transitioning from Submitted to Succeed, confirming that the scheduled scaling took effect.

Key Specification Fields

scaleTargetRef

identifies the workload to be scaled (e.g., a Deployment). The jobs array defines cron‑style schedules ( schedule) and the desired replica count ( targetSize) for each scaling action.

Reference

GitHub repository (open source): https://github.com/AliyunContainerService/kubernetes-cronhpa-controller