YOW2018 - Events and Commands: Developing Asynchronous Microservices
5 likes3,394 views
The document discusses the implementation of transactions and queries within a microservice architecture, emphasizing the use of asynchronous messaging. It covers managing transactions through sagas, the complexities of querying across services, and different patterns like CQRS and orchestration versus choreography. Additionally, it highlights the benefits and drawbacks of various approaches, including event sourcing and the transactional outbox pattern for managing data consistency and communication between services.
YOW2018 - Events and Commands: Developing Asynchronous Microservices
- 1. @crichardson Events and Commands: Developing Asynchronous Microservices Chris Richardson Founder of Eventuate.io Founder of the original CloudFoundry.com Author of POJOs in Action and Microservices Patterns @crichardson [email protected] http://learn.microservices.io Copyright © 2018 Chris Richardson Consulting, Inc. All rights reserved
- 2. @crichardson Presentation goal Implementing transactions and queries in a microservice architecture using asynchronous messaging
- 3. @crichardson Presentation goal Microservices > REST Microservices > Events Microservices != Event Sourcing Apache Kafka != Event Store
- 6. @crichardson Agenda Transactions, queries and microservices Managing transactions with sagas Implementing queries with CQRS Implementing transactional messaging
- 7. The microservice architecture structures an application as a set of loosely coupled services
- 8. @crichardson API Service = independently deployable component Operations Event Publisher Commands Queries Synchronous REST/gRPC Asynchronous Messaging Events Event Subscriber API Client Invokes Operations Events Service Database
- 9. @crichardson Microservices enable continuous delivery/deployment Process: Continuous delivery/deployment Organization: Small, agile, autonomous, cross functional teams Architecture: Microservice architecture Enables Enables Enables Successful Software Development Services = testability and deployability Teams own services
- 10. @crichardson Let’s imagine that you are building an online store API createCustomer(creditLimit) createOrder(customerId, orderTotal) findOrdersForCustomer(customerId) findRecentCustomers() Order Management Customer Management REST API …
- 11. @crichardson Order Service createCustomer() createOrder() findOrdersForCustomer() findRecentCustomers() API Gateway createCustomer() createOrder() Order DatabaseCustomer Database Order tableCustomer table REST API REST API Essential for loose coupling Must reserve customer’s credit Retrieve data from both services Customer Service REST API availableCredit …. orderTotal ….
- 12. @crichardson No ACID transactions that span services BEGIN TRANSACTION … SELECT ORDER_TOTAL FROM ORDERS WHERE CUSTOMER_ID = ? … SELECT CREDIT_LIMIT FROM CUSTOMERS WHERE CUSTOMER_ID = ? … INSERT INTO ORDERS … … COMMIT TRANSACTION Private to the Order Service Private to the Customer Service Distributed transactions
- 13. @crichardson Querying across services is not straightforward SELECT * FROM CUSTOMER c, ORDER o WHERE c.id = o.ID AND c.id = ? Private to Customer Service Private to Order Service Find customer and their orders
- 14. @crichardson Agenda Transactions, queries and microservices Managing transactions with sagas Implementing queries with CQRS Implementing transactional messaging
- 15. @crichardson From a 1987 paper
- 16. @crichardson Saga Use Sagas instead of 2PC Distributed transaction Service A Service B Service A Local transaction Service B Local transaction Service C Local transaction X Service C https://microservices.io/patterns/data/saga.html
- 17. @crichardson Order Service Create Order Saga Local transaction Order state=PENDING createOrder() Customer Service Local transaction Customer reserveCredit() Order Service Local transaction Order state=APPROVED approve order() createOrder() Initiates saga
- 18. @crichardson Saga design challenges API design Synchronous REST API initiates asynchronous saga When to send back a response? Rollback compensating transactions Sagas are ACD - No I Sagas are interleaved anomalies, such as lost updates Must use countermeasures https://www.slideshare.net/chris.e.richardson/saturn-2018-managing-data-consistency-in-a-microservice-architecture-using-sagas
- 19. How do the saga participants communicate? Synchronous communication, e.g. REST = temporal coupling Client and server need to be both available Customer Service fails retry provided it’s idempotent Order Service fails Oops Order Service createOrder() REST Customer Service reserveCredit() REST
- 20. @crichardson Collaboration using asynchronous, broker-based messaging Order Service Customer Service …. Message broker
- 21. @crichardson About the message broker At least once delivery Ensures a saga completes when its participants are temporarily unavailable Ordered delivery Mechanism for scaling consumers that preserves ordering e.g. Apache Kafka consumer group ActiveMQ message group …
- 22. @crichardson Saga step = a transaction local to a service Service Database Message Broker update publish message/event How to make atomic without 2PC?
- 23. @crichardson How to sequence the saga transactions? After the completion of transaction Ti “something” must decide what step to execute next Success: which T(i+1) - branching Failure: C(i - 1)
- 24. @crichardson Choreography: distributed decision making vs. Orchestration: centralized decision making
- 25. @crichardson Transactions, queries and microservices Managing transactions with sagas Implementing queries with CQRS Implementing transactional messaging Overview Choreography Orchestration Agenda
- 26. @crichardson Message broker Choreography-based Create Order Saga Order created Credit Reserved Credit Limit Exceeded Create Order OR Customer availableCredit ... Order state total create() reserveCredit() approve()/ reject() Order events channel Customer events channel Order Service Customer Service
- 27. Benefits and drawbacks of choreography Benefits Simple, especially when using event sourcing Participants are loosely coupled Drawbacks Decentralized implementation - potentially difficult to understand Cyclic dependencies - services listen to each other’s events, e.g. Customer Service must know about all Order events that affect credit Overloads domain objects, e.g. Order and Customer know too much Events = indirect way to make something happen https://github.com/eventuate-examples/eventuate-examples-java-customers-and-orders
- 28. @crichardson Agenda Transactions, queries and microservices Managing transactions with sagas Implementing queries with CQRS Implementing transactional messaging Overview Choreography Orchestration
- 29. @crichardson Order Service Orchestration-based coordination using command messages Local transaction Order state=PENDING createOrder() Customer Service Local transaction Customer reserveCredit() Order Service Local transaction Order state=APPROVED approve order() createOrder() CreateOrderSaga InvokesInvokesInvokes
- 30. @crichardson A saga (orchestrator) is a persistent object that implements a state machine and invokes the participants
- 31. Saga orchestrator behavior On create: Invokes a saga participant Persists state in database Wait for a reply On reply: Load state from database Determine which saga participant to invoke next Invokes saga participant Updates its state Persists updated state Wait for a reply …
- 32. @crichardson Order Service CreateOrderSaga orchestrator Customer Service Create Order Customer creditLimit creditReservations ... Order state total… reserveCredit CreateOrder Saga OrderService create() create() approve() Credit Reserved Customer command channel Saga reply channel https://github.com/eventuate-tram/eventuate-tram-sagas-examples-customers-and-orders
- 33. Benefits and drawbacks of orchestration Benefits Centralized coordination logic is easier to understand Reduced coupling, e.g. Customer Service knows less. Simply has API for managing available credit. Reduces cyclic dependencies Drawbacks Risk of smart sagas directing dumb services
- 34. @crichardson Agenda Transactions, queries and microservices Managing transactions with sagas Implementing queries with CQRS Implementing transactional messaging
- 35. @crichardson Queries often retrieve data owned by multiple services
- 36. @crichardson API Composition pattern Customer Service Customer … Order Service Order … API Gateway findOrdersForCustomer(customerId) GET /customer/id GET /orders?customerId=id https://microservices.io/patterns/data/api-composition.html
- 37. @crichardson Find recent, valuable customers SELECT * FROM CUSTOMER c, ORDER o WHERE c.id = o.ID AND o.ORDER_TOTAL > 100000 AND o.STATE = 'SHIPPED' AND c.CREATION_DATE > ? Customer Service Order Service …. is even more difficult!
- 38. API Composition would be inefficient 1 + N strategy: Fetch recent customers Iterate through customers fetching their shipped orders Lots of round trips high-latency Alternative strategy: Fetch recent customers Fetch recent orders Join 2 roundtrips but potentially large datasets inefficient
- 39. @crichardson Using events to update a queryable replica = CQRS Order Service Customer Service Order events Customer events findCustomersAndOrders() Order events channel Customer events channel Customer Order View Service Replica View Database https://microservices.io/patterns/data/cqrs.html
- 40. @crichardson Persisting a customer and order history in MongoDB { "_id" : "0000014f9a45004b 0a00270000000000", "name" : "Fred", "creditLimit" : { "amount" : "2000" }, "orders" : { "0000014f9a450063 0a00270000000000" : { "state" : "APPROVED", "orderId" : "0000014f9a450063 0a00270000000000", "orderTotal" : { "amount" : "1234" } }, "0000014f9a450063 0a00270000000001" : { "state" : "REJECTED", "orderId" : "0000014f9a450063 0a00270000000001", "orderTotal" : { "amount" : "3000" } } } } Denormalized = efficient lookup Customer information Order information
- 41. @crichardson Query side data model Command Query Responsibility Segregation (CQRS) Command side data model Commands Aggregate Message broker or Event Store Events Queries (Materialized) View Events POST PUT DELETE GET
- 42. @crichardson Queries database (type) Command side POST PUT DELETE Aggregate Event Store/Message Broker Events Query side GET /customers/id MongoDB Query side GET /orders?text=xyz ElasticSearch Query side GET … Neo4j
- 43. @crichardson CQRS views are disposable Rebuild when needed from source of truth Using event sourcing (Conceptually) replay all events from beginning of time Using traditional persistence “ETL” from source of truth databases
- 44. @crichardson Handling replication lag Lag between updating command side and CQRS view Risk of showing stale data to user Either: Update UI/client-side model without querying Use updated aggregate version to “wait” for query view to be updated
- 45. @crichardson Agenda Transactions, queries and microservices Managing transactions with sagas Implementing queries with CQRS Implementing transactional messaging
- 46. @crichardson Messaging must be transactional Service Database Message Broker update publish How to make atomic without 2PC?
- 47. Publish to message broker first? Guarantees atomicity BUT Service can’t read its own writes Difficult to write business logic Service Database Message Broker update publish
- 48. @crichardson Option: Event sourcing = Event centric approach to business logic and persistence http://eventuate.io/
- 49. @crichardson Event sourcing: persists an object as a sequence of events Event table Entity type Event id Entity id Event data Order 902101 …OrderApproved Order 903101 …OrderShipped Event type Order 901101 …OrderCreated Order create() approve() ship() Event Store
- 50. @crichardson Replay events to recreate in memory state Order state apply(event) Event table Entity type Event id Entity id Event data Order 902101 …OrderApproved Order 903101 …OrderShipped Event type Order 901101 …OrderCreated Event Store Load events by ID and call apply() Instantiate with default constructor Event store = database
- 51. @crichardson FYI: Apache Kafka != event store
- 52. @crichardson Event sourcing guarantees: state change event is published Event table Entity type Event id Entity id Event data Order 902101 …OrderApproved Order 903101 …OrderShipped Event type Order 901101 …OrderCreated Event Store Customer Service Subscribe Event store = message broker
- 53. @crichardson Preserves history of domain objects Supports temporal queries Simplifies retroactive correction Built-in auditing Other benefits of event sourcing
- 54. @crichardson Unfamiliar programming model Evolving the schema of long-lived events Event store only supports PK-based access requires CQRS less consistent reads Drawbacks of event sourcing
- 55. @crichardson Good fit for choreography-based sagas BUT orchestration is more challenging Use event handler to translate event into command/reply message Drawbacks of event sourcing
- 56. @crichardson Option: Traditional persistence (JPA, MyBatis,…) + Transactional outbox pattern https://github.com/eventuate-tram/eventuate-tram-core
- 57. @crichardson Spring Data for JPA example Publish event Save order http://eventuate.io/exampleapps.html
- 58. @crichardson Transactional outbox pattern ACID transaction See BASE: An Acid Alternative, http://bit.ly/ebaybase DELETE ? Customer Service ORDER_ID CUSTOMER_ID TOTAL 99 CUSTOMER_CREDIT_RESERVATIONS table 101 1234 ID TYPE DATA DESTINATION MESSAGE table 84784 CreditReserved {…} … INSERT INSERT Message Publisher QUERY Message Broker Publish Local transaction reserveCredit() DomainEventPublisher
- 59. @crichardson Transaction log tailing Order Service Datastore ORDER table Transaction log Update Transaction log miner Message Broker Publish Changes MySQL binlog Postgres WAL AWS DynamoDB table streams MongoDB change streams http://eventuate.io/
- 60. @crichardson Polling the message table Simple Works for all databases BUT what about polling frequency MESSAGE Table Message Publisher Message Broker SELECT * FROM MESSAGE… UPDATE MESSAGE
- 61. @crichardson Summary Use asynchronous messaging to solve distributed data management problems Services publish events to implement choreography-based sagas queries using CQRS views Services send command/reply messages to implement orchestration-based sagas Services must atomically update state and send messages Event sourcing Transactional outbox