reactive-architecture

Architecture

schedule-connector

• Receive schedule request from HTTP Endpoint
• Validate request body and require fields before response
  • If the validation or Kafka sender fails return bad request
  • If the validation and kafka sender ok return accepted http code
• Endpoint response a request Id to get processed request back
• Produce "ScheduleRequested" event to Kafka topic

schedule-schema

• Define Schedule structure

schedule-command

• Consume "ScheduleRequested" event from Kafka
  • Kafka Consumer with auto commit = false
  • Only filled fields will be considered
  • Processing batch messages and commit only at the end
  • If the invalid schema: offset will be committed and message will be sent to a DLQ (TODO)
  • If the schedule has invalid business, the offset will be committed and message will be sent to a DLQ (TODO)
  • If it happens any unexpected error, the batch messages will not be commit
  • If it happens unexpected error in last message, the previous messages could be processed
• All schedule are persisted in Cassandra
  • Schedule and Customer persist in a single table
  • Table key is composed by dateTime, description and customer Document number
• After processed produce schedule event
  • ScheduleProcessed is produced with requestId
  • Schedule change topic is produced

Running environment

1. Start infrastructure (Kafka, Cassandra), execute docker-compose up
2. Create topics ./kafka-create-topics.sh
3. Create tables ./cassandra-create-tables.sh

Building applications

1. Build docker images ./build-app-docker-images.sh

Running applications

1. Start applications ./start-apps.sh
2. Starting websocket client to receive process responsedocker run -it --network=reactive-microservices solsson/websocat ws://schedule-connector:8080/schedules
3. Execute test to send schedules ./test.sh