Question

A company runs a customer-facing payment API on AWS Lambda behind API Gateway. Recent production deployments have twice caused elevated error rates that went unnoticed for several minutes, resulting in failed transactions before engineers manually rolled back. The team uses AWS CodePipeline and CodeDeploy for deployments. Leadership wants future deployments to automatically shift only a small portion of traffic to the new version first, monitor for errors, and roll back automatically if problems are detected — all with minimal custom code. Which approach best meets these requirements?

Accepted Answer

Configure CodeDeploy to deploy the Lambda function using a canary traffic-shifting configuration with Lambda aliases, and attach a CloudWatch alarm on the function's error metric as a rollback trigger.

Answer

Publish each new Lambda version and use a weighted Route 53 record to send 10% of traffic to the new API Gateway stage, then update the weights manually after monitoring dashboards.

Answer

Deploy the new Lambda version with an all-at-once CodeDeploy configuration and create an EventBridge rule that invokes a rollback Lambda when the CloudWatch error alarm fires.

Answer

Enable API Gateway stage canary deployments to split traffic, and write a custom CloudWatch Synthetics canary plus a Step Functions workflow to evaluate metrics and revert the stage if errors exceed a threshold.