Serverless Deploy — Deployment Strategies and CI/CD

# Deploy all functions and resources
sls deploy --stage production

# Deploy a single function (faster)
sls deploy function -f createUser

# Deploy only the function code without infrastructure changes
sls deploy function -f createUser --update-config

# Package without deploying
sls package --stage production

# deploy_process.py
# Understanding the deployment process

def simulate_deploy(stage="dev"):
    steps = [
        "1. Validate serverless.yml syntax",
        "2. Package function code and dependencies",
        "3. Upload to S3 deployment bucket",
        "4. Generate CloudFormation template",
        "5. Execute CloudFormation stack change set",
        "6. Update Lambda function code",
        "7. Update Lambda function configuration",
        "8. Deploy API Gateway stage",
    ]
    
    print(f"Deploying to {stage}...\n")
    for step in steps:
        print(f"  {step}")
    print(f"\n  Stack: my-service-{stage}")
    print(f"  API URL: https://abc123.execute-api.us-east-1.amazonaws.com/{stage}/")

simulate_deploy("production")

# canary_deploy.py
# Canary deployment simulation

class CanaryDeployment:
    def __init__(self, function_name):
        self.function = function_name
        self.current_version = 1
        self.new_version = None
    
    def deploy_canary(self, traffic_percent=5):
        self.new_version = self.current_version + 1
        print(f"Deploying v{self.new_version} to {self.function}")
        print(f"  Routing {traffic_percent}% of traffic to new version")
        print("  Monitoring errors and latency...")
        
        if traffic_percent <= 10:
            print("  -> Canary: No errors detected")
            self.promote()
        else:
            print("  -> Canary: Error rate exceeded threshold")
            self.rollback()
    
    def promote(self):
        print(f"  Promoting v{self.new_version} to 100% traffic")
        self.current_version = self.new_version
    
    def rollback(self):
        print(f"  Rolling back to v{self.current_version}")

deploy = CanaryDeployment("order-api")
deploy.deploy_canary(traffic_percent=5)

# rollback.py
# Rollback strategies

def simulate_rollback(function_name, deploy_id, reason):
    print(f"Rolling back {function_name}")
    print(f"  Failed deployment: {deploy_id}")
    print(f"  Reason: {reason}")
    print(f"  Reverting to previous version...")
    
    steps = [
        "Update alias to point to previous version",
        "Redeploy previous CloudFormation stack",
        "Restore API Gateway stage",
        "Clear any warmed execution environments",
    ]
    
    for s in steps:
        print(f"  -> {s}")
    
    print(f"  Rollback complete. Function restored to previous state.")

simulate_rollback("create-user", "deploy-42", "Error rate exceeded 5% threshold")

# .github/workflows/deploy.yml
name: Deploy Serverless
on:
  push:
    branches: [main]

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - uses: actions/setup-node@v3
        with:
          node-version: 18
      - run: npm ci
      - run: npm test
      - name: Deploy to staging
        run: npx sls deploy --stage staging
        env:
          AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
          AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}

import random

def deploy_pipeline():
    stages = {
        "package": {"status": "pending"},
        "deploy-staging": {"status": "pending"},
        "tests": {"status": "pending"},
        "approve": {"status": "pending"},
        "canary-prod": {"status": "pending"},
        "promote": {"status": "pending"},
    }
    
    for stage in stages:
        print(f"[{stage}] Running...")
        stages[stage]["status"] = "passed" if random.random() > 0.1 else "failed"
        
        if stages[stage]["status"] == "failed":
            print(f"[{stage}] FAILED - Initiating rollback")
            for prev in reversed(list(stages.keys())):
                if stages[prev]["status"] == "passed":
                    print(f"  Rolling back: {prev}")
                    stages[prev]["status"] = "rolled-back"
            return False
        
        print(f"[{stage}] OK")
    
    print("Deployment complete!")
    return True

deploy_pipeline()