node-group-scaler

Node Group Scaler Operator

A Kubernetes operator for managing automatic node group scaling based on various triggers and cloud providers.

Overview

The Node Group Scaler Operator is a Kubernetes operator that automates the scaling of node groups across different cloud providers. It provides a flexible and powerful way to manage your cluster’s infrastructure based on various triggers and conditions.

Features

• Automatic node group scaling based on cron schedules
• Definition of minimum and maximum node limits
• Namespace-based configuration
• AWS EKS node group support (with plans to support other cloud providers)
• Flexible trigger system (currently supporting cron schedules, with plans to add more trigger types)
• Support for multiple scaling policies per node group
• Graceful scaling with configurable cooldown periods
• Detailed metrics and logging for scaling operations

Installation

Prerequisites

• Kubernetes cluster 1.19+
• Helm 3.0+
• AWS CLI configured (for AWS EKS support)
• kubectl configured to communicate with your cluster

Using Helm

# Add the Helm repository
helm repo add node-group-scaler https://charts.nodegroupscaler.io
helm repo update

# Install the operator
helm install node-group-scaler node-group-scaler/node-group-scaler \
  --namespace node-group-scaler \
  --create-namespace

Manual Installation

kubectl apply -f https://raw.githubusercontent.com/joao208/node-group-scaler/main/deploy/manifests.yaml

Configuration

NodeGroupScaler CRD

apiVersion: scaling.nodegroupscaler.io/v1alpha1
kind: NodeGroupScalingPolicy
metadata:
  name: example-policy
  namespace: default
spec:
  nodeGroup:
    name: my-node-group
    provider: aws
    region: us-west-2
  scaling:
    minNodes: 2
    maxNodes: 10
  schedule:
    - cron: "0 9 * * 1-5" # Scale up at 9 AM on weekdays
      targetNodes: 8
    - cron: "0 18 * * 1-5" # Scale down at 6 PM on weekdays
      targetNodes: 2
  cooldownPeriod: 300 # 5 minutes between scaling operations

Usage Examples

Basic Scaling Policy

apiVersion: scaling.nodegroupscaler.io/v1alpha1
kind: NodeGroupScalingPolicy
metadata:
  name: basic-scaling
spec:
  nodeGroup:
    name: production-nodes
    provider: aws
  scaling:
    minNodes: 3
    maxNodes: 15
  schedule:
    - cron: "0 8 * * *"
      targetNodes: 10

Multiple Scaling Policies

apiVersion: scaling.nodegroupscaler.io/v1alpha1
kind: NodeGroupScalingPolicy
metadata:
  name: complex-scaling
spec:
  nodeGroup:
    name: production-nodes
    provider: aws
  scaling:
    minNodes: 3
    maxNodes: 20
  schedule:
    - cron: "0 8 * * 1-5"
      targetNodes: 15
      description: "Weekday morning scale up"
    - cron: "0 18 * * 1-5"
      targetNodes: 5
      description: "Weekday evening scale down"
    - cron: "0 10 * * 0,6"
      targetNodes: 8
      description: "Weekend morning scale up"

Monitoring and Logging

The operator provides detailed metrics and logs for all scaling operations:

• Detailed logs with scaling decisions
• Status updates in the NodeGroupScalingPolicy resource

Troubleshooting

Common issues and their solutions:

1. Scaling not occurring

   • Check the operator logs
   • Verify cron schedule syntax
   • Ensure node group exists and is accessible

2. Permission issues

   • Verify IAM roles and permissions
   • Check service account configuration

3. Scaling limits reached

   • Review min/max node configurations
   • Check for any quota limitations

Roadmap

• Support for additional cloud providers:
  • Google Cloud Platform (GKE)
  • Azure (AKS)
  • DigitalOcean
• Additional trigger types:
  • Metrics-based scaling (CPU, Memory, Custom metrics)
  • Event-based scaling
  • Time-based scaling with timezone support
  • Manual scaling triggers
• Enhanced features:
  • Cost optimization recommendations
  • Predictive scaling based on historical patterns
  • Multi-cluster support
  • Webhook integration for custom scaling logic

Contributing

We welcome contributions! Please see our Contributing Guide for details on how to submit pull requests, report issues, and contribute to the project.

Development Setup

1. Fork the repository
2. Clone your fork
3. Install dependencies:

   make setup
   

Support

• GitHub Issues: Report a bug
• Documentation: Full documentation
• Community: Join our Slack channel