We’ve noticed that our architecture is not prepared to scale the Ingestion/Consumption layer yet, something that we will need to solve in another workshop.

Considering that, for the time being, we will only have one (1) shard for Kinesis Data Streams, find the limits for this architecture.


  • What are the limits for Kinesis Streams for one shard?
  • How is the integration of Lambda and Kinesis affected if we scale up to more users? Do we need to worry about the “batch size” for the Lambda-Kinesis integration? Or do we just need to fine tune it?
  • How will DynamoDB scale with increased traffic? What will that look like? Will we need to set that up?