Kubernetes Schedule Scaler

This Application/ Kubernetes controller is used to schedule scale deployments in the cluster based in annotations. The controller can work in conjunction with hpa. if hpa is configured the controller can adjust minReplicas and maxReplicas. At the moment it supports reading the scaling definitions from directly in the annotations.

Why fork?

A less common resource(stackset) was targeted and we wanted to remove it.
We wanted to remove the ability to save to S3 as well because it's hard to use.

Usage

install

https://github.com/chatwork/charts/tree/master/kube-schedule-scaler#tldr

deployment

Add the annotation to either your Deployment.

  annotations:
    zalando.org/schedule-actions: '[{"schedule": "10 18 * * *", "replicas": "3"}]'

HPA

  annotations:
    zalando.org/schedule-actions: '[{"schedule": "10 18 * * *", "minReplicas": "3"}]'

Available Fields

The following fields are available

schedule - Typical crontab format
replicas - the number of replicas to scale to
minReplicas - in combination with an hpa will adjust the minReplicas else be ignored
maxReplicas - in combination with an hpa will adjust the maxReplicas else be ignored

Deployment Example

kind: Deployment
metadata:
  name: nginx-deployment
  labels:
    application: nginx-deployment
  annotations:
    zalando.org/schedule-actions: |
      [
        {"schedule": "30 4 * * 1,2,3,4,5", "replicas": "{{{HIGH_LOAD_REPLICAS}}}"},
        {"schedule": "00 8 * * 1,2,3,4,5", "replicas": "{{{REPLICAS}}}"}
      ]

HPA Example

apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  annotations:
    zalando.org/schedule-actions: |
      [
        {"schedule": "15 4 * * *", "minReplicas": "1"},
        {"schedule": "12 4 * * *", "maxReplicas": "20"},
        {"schedule": "10 4 * * *", "maxReplicas": "15", "minReplicas": "2"}
      ]
  name: php-apache
spec:
  maxReplicas: 20
  minReplicas: 1
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: php-apache
  targetCPUUtilizationPercentage: 50

autoscaling/v2beta2

apiVersion: autoscaling/v2beta2
kind: HorizontalPodAutoscaler
metadata:
  name: php-apache
  annotations:
    zalando.org/schedule-actions: |
      [
        {"schedule": "20 7 * * *", "maxReplicas": "15", "minReplicas": "2"}
      ]
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: php-apache
  minReplicas: 1
  maxReplicas: 7
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 50

chatwork / kube-schedule-scaler