Serverless Refactoring And Modernization — AWS Developer (DVA-C02)
Monolith-as-Lambda Hits the 15-Minute Ceiling
Lambda enforces a hard 15-minute maximum execution duration. Any path inside a monolith that needs longer fails at that limit, regardless of whether the rest of the application fits. Candidates read 'move to serverless' and package the application as a single Lambda function, which satisfies the runtime label while ignoring this constraint. Real decomposition means identifying which discrete operations complete within the window, separating stateless compute from stateful storage, and routing work between units with SQS or EventBridge. Per-function IAM roles fall out of that split; each function gets only the permissions its own code needs.
What This Pattern Tests
The exam describes a workload and tests whether serverless is the right modernization path. Lambda works for event-driven, short-duration workloads (API handlers, file processors, stream consumers). It fails for long-running processes (>15 min), workloads needing persistent connections (WebSockets need API Gateway WebSocket APIs), or high-throughput steady-state (Lambda per-invocation cost exceeds Fargate at ~1M requests/day). Lambda in a VPC adds cold start latency (VPC ENI creation) unless you use provisioned concurrency ($$$). Step Functions orchestrate multi-step workflows that exceed Lambda timeout. The trap is assuming Lambda is always the modernization answer.
Decision Axis
Workload duration, traffic pattern, and connection requirements determine whether serverless fits or a container is cheaper and simpler.
Associated Traps
More Top Traps on This Exam
Decision Rules
Choose reserved concurrency over provisioned concurrency to enforce an absolute cap on concurrent Lambda executions that simultaneously protects the downstream DynamoDB write capacity and prevents this function from consuming shared account concurrency budget.
Whether to cap Lambda throughput via reserved concurrency — which simultaneously limits concurrent writes to DynamoDB AND isolates account concurrency — or to increase DynamoDB write capacity to absorb the load, which removes the downstream bottleneck but leaves Lambda free to exhaust the shared account concurrency pool.
Publish the shared library as a Lambda Layer artifact (AWS::Serverless::LayerVersion) staged in S3 via SAM rather than bundling the 45 MB library inside each function's individual deployment ZIP.
Domain Coverage
Difficulty Breakdown