Monitoring và tối ưu chi phí

Cost Optimization Framework

Level 1: Basic Monitoring (Ngày 1-7)

Setup cơ bản:

1. AWS Budgets

   • Tạo các Budget để đảm bảo sử dụng “AWS an toàn” và có biện pháp xử lý kịp thời
     • Budget 1: $50/month (Alert at 80%)
     • Budget 2: $25/month (Alert at 50%)
     • Budget 3: $10/day (Alert at 100%)

2. Billing Alerts

   • Tạo các CloudWatch Alarms:
     • $25 threshold → Email warning
     • $50 threshold → Email + SMS
     • $75 threshold → Email + SMS + Slack

3. Cost Explorer

   • Enable daily cost reports
   • Set up service-level breakdown
   • Monitor top 5 cost drivers

Level 2: Advanced Analytics (Tuần 2-4)

Deep Dive Analysis:

1. Custom CloudWatch Metrics

   import boto3
   
   def publish_cost_metrics():
       ce = boto3.client('ce')
       cw = boto3.client('cloudwatch')
   
       # Get daily cost
       response = ce.get_cost_and_usage(
           TimePeriod={
               'Start': '2025-01-01',
               'End': '2025-01-02'
           },
           Granularity='DAILY',
           Metrics=['BlendedCost']
       )
   
       cost = float(response['ResultsByTime'][0]['Total']['BlendedCost']['Amount'])
   
       # Publish to CloudWatch
       cw.put_metric_data(
           Namespace='AWS/Billing/Custom',
           MetricData=[
               {
                   'MetricName': 'DailyCost',
                   'Value': cost,
                   'Unit': 'None'
               }
           ]
       )
   

2. Resource Tagging Strategy

   Mandatory Tags:
   - Project: project-name
   - Environment: dev/staging/prod
   - Owner: email@domain.com
   - CostCenter: department
   - AutoShutdown: true/false
   - CreatedDate: YYYY-MM-DD
   

Emergency Cost Control

Critical Actions (>$150 spent)

1. Immediate Resource Audit

   # Find most expensive resources
   aws ce get-cost-and-usage \
     --time-period Start=2025-01-01,End=2025-01-31 \
     --granularity MONTHLY \
     --metrics BlendedCost \
     --group-by Type=DIMENSION,Key=SERVICE
   

2. Emergency Shutdown Protocol

   # Stop all non-critical instances
   aws ec2 stop-instances --instance-ids $(
     aws ec2 describe-instances \
       --filters "Name=tag:Critical,Values=false" \
       --query 'Reservations[].Instances[?State.Name==`running`].InstanceId' \
       --output text
   )