Access CloudWatch Metrics:

1. Sign in to AWS Console
2. Go to Lambda → Your function (mlops-service-lambda)
3. Click the "Monitor" tab
4. You'll see automatic metrics graphs:

Access CloudWatch Logs:

1. In Lambda "Monitor" tab, click "View CloudWatch logs"
2. You'll see log streams - one per Lambda execution environment
3. Click on the latest log stream
4. View your application logs:

Example Log Output:

START RequestId: abc123 Version: $LATEST
2024-01-15T10:30:00.000Z INFO Received request to /health
2024-01-15T10:30:00.050Z INFO Database connection healthy
2024-01-15T10:30:00.100Z INFO Prometheus metrics updated
END RequestId: abc123
REPORT Duration: 150.00 ms Memory: 128 MB Max Memory: 89 MB

Generate some traffic to your Lambda:

# Replace with your API Gateway URL from Lab 8
API_URL="https://YOUR_API_GATEWAY_URL"

# Send 10 test requests
for i in {1..10}; do
  curl -s $API_URL/health
  sleep 1
done

Then refresh CloudWatch to see:

• Invocations count increased by 10
• Duration showing ~100-200ms per request
• First request may show higher duration (cold start)
• Logs showing your requests

Create SNS Topic:

1. Go to AWS Console → Search "SNS" → Open SNS
2. Click "Topics" in left sidebar → "Create topic"
3. Configure:
   • Type: Standard
   • Name: mlops-lambda-alerts
4. Click "Create topic"

Create Email Subscription:

1. In the topic page, click "Create subscription"
2. Configure:
   • Protocol: Email
   • Endpoint: Your email address
3. Click "Create subscription"
4. Check your email for confirmation link and click it

Create CloudWatch Alarm:

1. Go to CloudWatch → Alarms → "Create alarm"
2. Click "Select metric"
3. Navigate: Lambda → By Function Name
4. Select your function (mlops-service-lambda)
5. Check the Errors metric → "Select metric"
6. Configure alarm conditions:
   • Statistic: Sum
   • Period: 5 minutes
   • Threshold type: Static
   • Whenever Errors is: Greater than 5
7. Click "Next"

Configure notifications:

1. Select "In alarm" state trigger
2. Select your SNS topic: mlops-lambda-alerts
3. Click "Next"

Name the alarm:

1. Alarm name: mlops-lambda-high-errors
2. Description: "Alert when Lambda errors exceed 5 in 5 minutes"
3. Click "Next" → Review → "Create alarm"

Update mlops-service/requirements.txt:

# Existing dependencies remain...

# Additional utilities
python-dotenv==1.0.0
psutil==5.9.8  # NEW: For system monitoring

Install the new dependency:

cd mlops-service
pip install psutil==5.9.8

Start your local Flask service:

cd mlops-service
python app.py

Test the detailed health endpoint:

curl http://localhost:5001/health/detailed | jq

Expected Output:

{
  "status": "healthy",
  "service": "mlops-service-prometheus",
  "version": "1.0.0",
  "timestamp": "2024-01-15T10:30:00.000000",
  "environment": "development",
  "uptime": {
    "seconds": 3600.5,
    "hours": 1.0,
    "days": 0.04,
    "started_at": "2024-01-15T09:30:00.000000"
  },
  "system": {
    "memory_usage_mb": 125.4,
    "cpu_percent": 2.5,
    "process_id": 12345
  },
  "requests": {
    "total": 150,
    "failed": 3,
    "error_rate_percent": 2.0,
    "last_request": "2024-01-15T10:29:00.000000"
  },
  "health_checks": {
    "database_connected": true,
    "prometheus_enabled": true,
    "metrics_tracking": "active"
  }
}

Open the dashboard:

open http://localhost:5001/