Marcia Villalba

Getting started with Orchestration and Choreography in Distributed Applications by Marcia Villalba

Communication between different parts of your distributed application can become very complex if you don’t have good architecture practices.

In this session, you will learn how and when to apply two architectural patterns in your distributed applications: orchestration and choreography. In addition, you will learn how to orchestrate complex workflows with state machines and how an event bus can help you choreograph micro services.

And you will also get good practices on how to use them to make your application scale and be maintainable.

Talk Questions

  • Question 157
    how do you recommend to anonymize the data of the unprocessed events when an anonimization process is triggered?
  • Question 159
    Idempotency ID you think is compatible with correlation ID? Complementary? Better? What you think?
  • Question 155
    EventBridge vs Kinesis vs Kafka?
  • Question 153
    When to use a custom workflow orchestration and when to use a platform like airflow? Retries, observability, ease of use, and power wise
  • Question 160
    Could you give an example of switching from ordered events to a design with no need for consumers to keep track of the order?
  • Question 154
    Which is the threshold (in terms of millions events per day) where Eventbrite starts to be less convenient than a Kafka cluster (running in eks or ec2 instances)?
  • Question 158
    Is there a nice way of enforcing a contract between microsevices?
  • Question 161
    One know characteristics of SQS (when it's not an FIFO) is the duplication of messages. Eventbridge can guarantee the deduplication of the message?
  • Question 162
    Saga: what if compensation action is also failing? Compensation on compensating action?
  • Question 163
    How do you test complex flows locally?
  • Question 156
    in the EDA, how do you bring the status of the request you your user? Several tasks ara happening in parallel