Stories

Problems from a production platform I work on: what I tried first, what broke, and what I would do again.

• Why I became more careful with DynamoDB Streams

  Stream-triggered processing looked clean until duplicates, batch retries and silently swallowed errors showed up. Moving website scraping to SQS made failure handling explicit.

  • DynamoDB Streams
  • SQS
  • Lambda

• Scheduled email campaigns, end to end

  One durable task and one EventBridge Scheduler schedule per email, a reconciler for drift, and live delivery status in the UI.

  • EventBridge Scheduler
  • Lambda
  • DynamoDB
  • React

• Why auto-generated CRUD was not enough

  Where custom AppSync mutations backed by Lambda took over from generated CRUD: batch commands, guarded deletes and payload limits.

  • AppSync
  • GraphQL
  • Lambda
  • Amplify Gen 2

• How webhooks completed the async task loop

  External services report job status through webhooks. The handler treats them as notifications, fetches results itself and acknowledges duplicates idempotently.

  • API Gateway
  • Webhooks
  • Lambda
  • DynamoDB

• I forgot try/catch in a Lambda. AWS did not forget.

  Unhandled errors are not just code errors, they are infrastructure behavior: how synchronous and asynchronous invocations retry, and what I do about it.

  • Lambda
  • SQS
  • EventBridge