Cost optimization

# Detect cost anomalies
clanker cost anomalies

# This will show:
# - Idle EC2 instances
# - Unattached EBS volumes
# - Old snapshots
# - Unused RDS instances
# - Underutilized resources

========================================
Cost Anomalies - Potential Savings: $579/month
========================================

🚨 High Impact:

1. Idle EC2: i-0a1b2c3d4e5f6 (dev-test-1)
   - Running 24/7 with 0.2% avg CPU
   - Last login: 14 days ago
   - Monthly waste: $125
   - Fix: Stop or terminate

2. Unused RDS: dev-playground-db
   - No connections in 30 days
   - Monthly cost: $178
   - Fix: Take snapshot, delete

3. Unattached EBS: 8 volumes, 1.2 TB
   - Monthly cost: $120
   - Fix: Delete unused volumes

# Stop idle EC2 instances
clanker ask --maker "stop EC2 instance i-0a1b2c3d4e5f6"
clanker ask --apply < plan.json

# Delete unused volumes
clanker ask --maker --destroyer "delete unattached EBS volumes older than 30 days"

# Delete unused RDS
clanker ask --maker --destroyer "take final snapshot of dev-playground-db then delete it"

# Check CPU utilization
clanker ask "Show me EC2 instances with low CPU usage over the last 7 days"

# Example output:
# prod-app-1 (c5.2xlarge): avg 12% CPU, max 23%
# api-server (m5.xlarge): avg 8% CPU, max 15%

# Change instance type
clanker ask --maker "change prod-app-1 from c5.2xlarge to c5.xlarge"

# The plan will:
# 1. Stop the instance
# 2. Modify instance type
# 3. Start the instance

# Review and apply
clanker ask --apply < plan.json

c5.2xlarge: $0.34/hour × 730 hours = $248/month
c5.xlarge:  $0.17/hour × 730 hours = $124/month
Savings: $124/month per instance

clanker ask "Show me RDS instances with low CPU and connection count"

# staging-db: db.t3.large, avg 5% CPU, max 3 connections
# dev-db: db.t3.medium, avg 3% CPU, max 1 connection

# Create snapshot first (safety)
clanker ask --maker "create snapshot of staging-db"

# Modify instance class
clanker ask --maker "change staging-db from db.t3.large to db.t3.small"

# Apply immediately or during maintenance window
clanker ask --apply < plan.json

db.t3.large: $0.136/hour × 730 hours = $99.28/month
db.t3.small: $0.034/hour × 730 hours = $24.82/month
Savings: $74.46/month

# Check memory usage
clanker ask "Show me Lambda functions with max memory usage vs. allocated"

# Example:
# image-processor: allocated 1024 MB, max used 567 MB
# api-handler: allocated 512 MB, max used 128 MB

# Reduce memory (also reduces CPU)
clanker ask --maker "change image-processor memory from 1024 MB to 768 MB"
clanker ask --maker "change api-handler memory from 512 MB to 256 MB"

clanker ask --apply < plan.json

# Create lifecycle policy
clanker ask --maker "create S3 lifecycle policy for logs-archive bucket:
- Move to Intelligent-Tiering after 30 days
- Move to Glacier after 90 days
- Delete after 365 days"

clanker ask --apply < plan.json

S3 Standard:            $0.023/GB/month
S3 Intelligent-Tiering: $0.0125/GB/month (infrequent access)
S3 Glacier:             $0.004/GB/month

For 1 TB of old logs:
Before: $23/month
After:  $4/month (Glacier)
Savings: $19/month = $228/year

# gp3 is 20% cheaper than gp2 with same performance
clanker ask "Show me gp2 EBS volumes"

# Convert to gp3
clanker ask --maker "convert all gp2 volumes to gp3"

gp2: $0.10/GB/month
gp3: $0.08/GB/month

For 500 GB:
Savings: $10/month = $120/year

# Find old snapshots
clanker ask "Show me EBS snapshots older than 90 days"

# Delete (after verifying they're not needed)
clanker ask --maker --destroyer "delete EBS snapshots older than 90 days for terminated instances"

# Find steady-state workloads
clanker ask "Show me EC2 instances running 24/7 for the last 30 days"

# Check RDS uptime
clanker ask "Show me RDS instances with high uptime"

On-Demand c5.xlarge: $0.17/hour
1-year RI (no upfront): $0.112/hour (34% savings)
3-year RI (all upfront): $0.091/hour (46% savings)

For 3 instances running 24/7:
On-Demand: $3,733/year
1-year RI: $2,461/year
Savings: $1,272/year (34%)

# Check pod resource usage
clanker k8s stats pods -A --sort-by memory

# Find pods using much less than requested
clanker k8s ask "show me pods using less than 50% of requested CPU or memory"

# Reduce requests and limits
kubectl set resources deployment my-app \
  --requests=cpu=100m,memory=256Mi \
  --limits=cpu=500m,memory=512Mi

# Or use VPA (Vertical Pod Autoscaler)
kubectl apply -f - <<EOF
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"
EOF

# EKS cluster autoscaler
helm install cluster-autoscaler autoscaler/cluster-autoscaler \
  --set autoDiscovery.clusterName=my-cluster \
  --set awsRegion=us-east-1

# Scale down to 0 during off-hours (dev clusters)
kubectl scale deployment my-app --replicas=0

# Use CronJobs to automate
kubectl create cronjob scale-down \
  --schedule="0 18 * * 1-5" \
  --image=bitnami/kubectl \
  -- kubectl scale deployment my-app --replicas=0

kubectl create cronjob scale-up \
  --schedule="0 8 * * 1-5" \
  --image=bitnami/kubectl \
  -- kubectl scale deployment my-app --replicas=3

# Add spot node group to EKS
clanker ask --maker "add spot instance node group to my-cluster with:
- Instance types: t3.medium, t3.large
- Min: 1, Max: 10, Desired: 2
- Spot allocation strategy: lowest-price"

# Savings: 60-90% compared to on-demand

apiVersion: apps/v1
kind: Deployment
metadata:
  name: batch-processor
spec:
  template:
    spec:
      nodeSelector:
        node.kubernetes.io/instance-type: spot
      tolerations:
      - key: "spot"
        operator: "Exists"
        effect: "NoSchedule"

#!/bin/bash
# stop-dev-resources.sh (run at 6pm weekdays)

# Stop EC2 instances with Environment=dev tag
for instance in $(aws ec2 describe-instances \
  --filters "Name=tag:Environment,Values=dev" \
  --query "Reservations[].Instances[].InstanceId" \
  --output text); do
  aws ec2 stop-instances --instance-ids $instance
  echo "Stopped $instance"
done

# Stop RDS instances
for db in $(aws rds describe-db-instances \
  --query "DBInstances[?TagList[?Key=='Environment' && Value=='dev']].DBInstanceIdentifier" \
  --output text); do
  aws rds stop-db-instance --db-instance-identifier $db
  echo "Stopped $db"
done

echo "Dev resources stopped. Savings: ~$200/night"

#!/bin/bash
# start-dev-resources.sh (run at 8am weekdays)

for instance in $(aws ec2 describe-instances \
  --filters "Name=tag:Environment,Values=dev" \
  --query "Reservations[].Instances[].InstanceId" \
  --output text); do
  aws ec2 start-instances --instance-ids $instance
done

for db in $(aws rds describe-db-instances \
  --query "DBInstances[?TagList[?Key=='Environment' && Value=='dev']].DBInstanceIdentifier" \
  --output text); do
  aws rds start-db-instance --db-instance-identifier $db
done

# Add to crontab
0 18 * * 1-5 /usr/local/bin/stop-dev-resources.sh
0 8 * * 1-5 /usr/local/bin/start-dev-resources.sh

# Savings: 14 hours/day × 5 days = 70 hours/week
# For $200/week resources: $700/month savings

# S3 VPC endpoint (free)
clanker ask --maker "create VPC endpoint for S3 in my-vpc"

# DynamoDB VPC endpoint (free)
clanker ask --maker "create VPC endpoint for DynamoDB in my-vpc"

# Savings: $0.045/GB NAT Gateway processing
# For 500 GB/month: $22.50 savings

# Instead of serving from S3 directly
clanker ask --maker "create CloudFront distribution for S3 bucket my-assets"

# Savings on data transfer:
# S3: $0.09/GB
# CloudFront: $0.085/GB (plus caching reduces origin requests)

# AWS Budget
clanker ask --maker "create AWS Budget with:
- Monthly limit: $3000
- Alert at 80% ($2400)
- Alert at 100% ($3000)
- Email: [email protected]"

# CloudWatch billing alarm
clanker ask --maker "create CloudWatch alarm when estimated charges exceed $3000"

#!/bin/bash
# weekly-cost-review.sh

echo "=== Weekly Cost Review ==="
echo ""

echo "This week's spending:"
clanker cost summary --start $(date -d '7 days ago' +%Y-%m-%d)

echo ""
echo "Cost anomalies:"
clanker cost anomalies

echo ""
echo "Forecast for month:"
clanker cost forecast

echo ""
echo "Action items:"
echo "1. Review anomalies above and stop/delete unused resources"
echo "2. Check for right-sizing opportunities"
echo "3. Verify dev resources are stopped during off-hours"