DynamoDB and Apache Kafka are free.
AWS key-value and document database that scales without a server to patch.
Try itNo servers to run
Scales with traffic
Fits the rest of AWS
Awkward when you need joins
Easy to get the data model wrong
Tied to AWS
A log that services publish events to, and other services read later.
Try itProducers do not wait on every downstream service
You can replay events after a bug
It holds up once you actually have a stream of events
Running a cluster is a lot for one app
A database is the simpler choice until you need a stream