Hi, thanks and sorry for the delay :) In general, we try to avoid using lambdas. The company has around 350 microservices running in some k8s clusters, so we already have some default tooling to those. We only use lambdas in very specific cases when the AWS lambda plug and play system is way easier than using the SDK. Like when we need to repass some data from an SQS to a ElasticSearch or repass DynamoStream data to Kinesis.