Soru

Zorluk: ZorModernizing Workloads with Serverless Architectures (Lambda and API Gateway)

A healthcare provider is modernizing its patient portal backend by refactoring a legacy monolithic API into a serverless architecture. The new design uses Amazon API Gateway and AWS Lambda functions that query a PostgreSQL database in a private subnet. The architecture must satisfy the following requirements:

1. The API must be accessible only from the on-premises network via an existing AWS Direct Connect connection.
2. The database must be protected from connection exhaustion during sudden morning login spikes, while ensuring Lambda functions remain highly available.
3. A single high-volume notification function must not exhaust the regional concurrency limit and disrupt other critical portal operations.
4. All application logs must be encrypted using a KMS Customer Managed Key (CMK) that is shared with a centralized security account.

Which TWO configurations should the Solutions Architect implement to meet these requirements?

  1. Create a Private API in Amazon API Gateway, establish an interface VPC endpoint for API Gateway in the private subnets of the VPC, and apply an API Gateway resource policy that allows access only from the VPC endpoint.Cevap
  2. Deploy an Amazon RDS Proxy in the private subnets of the VPC to manage database connections, and configure a reserved concurrency limit on the high-volume notification Lambda function.Cevap
  3. C
    Use the default AWS managed key (aws/logs) to encrypt the Amazon CloudWatch log groups, and modify the key policy to grant the centralized security account cross-account decryption permissions.
  4. D
    Configure the private subnets where the Lambda functions are located to route outbound traffic through a single NAT Gateway deployed in a single public subnet to control egress costs, and rely on the default regional concurrency limit.
  5. E
    Enable provisioned concurrency on all Lambda functions to guarantee start times during peak morning spikes, and configure an Application Load Balancer to act as the entry point to route traffic directly to the private Lambda functions.

Cevap

Create a Private API in Amazon API Gateway, establish an interface VPC endpoint for API Gateway in the private subnets of the VPC, and apply an API Gateway resource policy that allows access only from the VPC endpoint. Additionally, deploy an Amazon RDS Proxy in the private subnets of the VPC to manage database connections, and configure a reserved concurrency limit on the high-volume notification Lambda function.
To establish a private API connection over Direct Connect, a Private API Gateway integrated with an interface VPC endpoint (VPCE) is required, secured by an API Gateway resource policy that restricts access to the VPC endpoint. To manage relational database connections effectively and prevent connection pool exhaustion from scaling serverless functions, Amazon RDS Proxy is deployed to pool database connections. Setting reserved concurrency on the high-volume function restricts its capacity, ensuring it cannot consume the entire account's regional concurrency pool and throttle other critical services.

Adım Adım Çözüm

1
Configure private API Gateway access
Create a Private API and attach an interface VPC endpoint (execute-api) inside the private subnets, restricting access with an API Gateway resource policy.
This allows secure on-premises access over AWS Direct Connect without routing traffic over the public internet.
2
Prevent database connection exhaustion
Deploy Amazon RDS Proxy in the target VPC private subnets.
RDS Proxy pools database connections, preventing the scaling Lambda functions from overwhelming the relational database's connection limits.
3
Isolate high-volume function concurrency
Set a reserved concurrency limit on the high-volume notification Lambda function.
Reserved concurrency acts as a ceiling for the function, ensuring it cannot consume the entire regional concurrency pool and throttle other critical microservices.

Anahtar Kavram

Serverless application modernization requiring secure private endpoints, database connection pooling, Lambda concurrency isolation, and compliance-aligned KMS key management.
Bu soruyu puanla