When Stateful Serverless met Planet-scale Event-sourcing

You are convinced your business requires Event-Sourcing, due to strict audit log requirements, traceability, easy anomaly & fraud detection, and who doesn't like time-traveling queries on their data after all. In today's world where everything is in the Cloud, You also want to utilize the true power of Cloud-Native components in Azure and design a highly scalable, secure and cost-effective solution.

Transcript

1. When Stateful Serverless met
   Planet-scale Event sourcing
   Based on true “Events”
   +
   

   View Slide

2. Answers have
   changed
   

   View Slide

3. Muhammad Sajid
   Cloud Solutions Architect with a passion for
   designing and developing cloud-native solutions.
   Interested in DDD, Distributed Event-Driven
   systems, BIG data, and IoT.
   @sajid_nazeer
   www.linkedin.com/in/musa
   

   View Slide

4. View Slide

5. Event sourcing is an approach
   in which changes to application
   state are saved as stream of
   events.
   

   View Slide

6. View Slide

7. ● A complete log of changes
   ● Traceability
   ● Death of cannot reproduce
   ● Scalability (with CQRS)
   

   View Slide

8. View Slide

9. View Slide

10. View Slide

11. View Slide

12. Entities are tiny services that
    communicate via messages and
    can have some state.
    

    View Slide

13. View Slide

14. View Slide

15. View Slide

16. View Slide

17. Durable Entities
    Addressable
    Always
    there
    State
    Execute
    serially
    

    View Slide

18. Glue code
    State
    

    View Slide

19. GET http(s):///runtime/webhooks/durabletask/entities/Toggle/1
    POST http(s):///runtime/webhooks/durabletask/entities/Toggle/1
    

    View Slide

20. Limitations?
    

    View Slide

21. View Slide

22. View Slide

23. View Slide

24. Entity-Entity
    communication
    happens via
    Signaling
    Beware of State
    mutation gotcha
    Not ideal for
    applications that
    require strict
    latency
    Request/
    Response
    pattern is only
    possible via
    Orchestrations
    

    View Slide

25. What is a
    Transactional
    Batch?
    Transactional Batch
    describes a group of
    operations that need
    to either succeed or
    fail as a batch
    

    View Slide

26. View Slide

27. View Slide

28. View Slide

29. If you talk to a man in a
    language he understands, that
    goes to his head.
    If you talk to him in his
    language, that goes to his heart
    - Nelson Mandela
    

    View Slide

30. Limitations?
    

    View Slide

31. Maximum
    operations per
    batch are 100
    Maximum
    payload size is
    2 MB
    Maximum
    Execution time is
    5 Seconds
    No Server-side
    Id generation
    

    View Slide

32. Demo
    

    View Slide

33. Takeaways
    ● Choose right partition key
    ● Beware of limitations of both Durable Entities and
    Transactional batch.
    ● Don’t settle for the right solutions . Find new ones, and
    keep learning.
    ● https://docs.microsoft.com/en-us/azure/azure-functions/durable/durable-fun
    ctions-entities?tabs=csharp
    ● https://devblogs.microsoft.com/cosmosdb/introducing-transactionalbatch-in-
    the-net-sdk/
    

    View Slide

34. Questions?
    

    View Slide

35. @sajid_nazeer
    www.linkedin.com/in/musa
    Molte grazie
    

    View Slide