Inter-service Communication

Communication Strategies

Arya Banking utilizes three primary communication patterns to ensure scalability, security, and loose coupling.

1. Synchronous Feign (REST)

Internal service-to-service calls are handled by OpenFeign. To secure these calls, we use the OAuth2 Client Credentials grant.

Machine-to-Machine (M2M) Flow

1. Source Service (e.g., Auth Service) triggers a Feign call.
2. OAuth2 Interceptor requests a machine-to-machine JWT from Keycloak using its own client-id and client-secret.
3. Keycloak returns a JWT with ROLE_INTERNAL_SERVICE.
4. Source Service injects Authorization: Bearer <JWT> into the outgoing request.
5. Target Service (e.g., User Service) validates the JWT and verifies the role.

Implementation Pattern

// OAuth2FeignConfig (Common Library pattern)
@Bean
public RequestInterceptor oauth2RequestInterceptor() {
    return requestTemplate -> {
        OAuth2AuthorizeRequest request = OAuth2AuthorizeRequest
            .withClientRegistrationId(clientRegistrationId).build();
        OAuth2AuthorizedClient client = authorizedClientManager.authorize(request);
        requestTemplate.header("Authorization", "Bearer " + client.getAccessToken().getTokenValue());
    };
}

2. API Flow Deep-Dive

User Registration Sync

When a user registers, the flow spans two services:

Account Locking (Login Failures)

When login fails multiple times, the Auth Service signals the User Service:

3. Asynchronous Events (Kafka) — UPDATED

State changes are propagated asynchronously using Apache Kafka and Avro Schemas.

Event Topics & Producers/Consumers

Producer: Auth Service User Registration

Producer: Auth Service Login Failure

Producer Logic (UserEventProducer in Auth Service)

The UserEventProducer in the Auth Service uses a typed KafkaTemplate<String, LoginFailedEvent> to send Avro-encoded records:

// Auth Service - UserEventProducer
kafkaTemplate.send(AUTH_FAILED_TOPIC, event.getUserId().toString(), event);

Consumer Logic (UserUpdateEventListener in Auth Service)

// Auth Service - UserUpdateEventListener
@KafkaListener(id = "user-update-event", topics = USER_UPDATE_TOPIC)
public void onUserUpdateEvent(OutboxKafkaEvent event) {
    UserCreateEvent userCreateEvent = GsonParser.fromJson(
        event.getPayload().toString(), UserCreateEvent.class);
    keyCloakService.onUserUpdateEvent(userCreateEvent);
}

Consumer Logic (UserEventListeners in User Service)

// User Service - UserEventListeners
@KafkaListener(id = "login-failed-event", topics = AUTH_FAILED_TOPIC)
public void onUserUpdateEvent(LoginFailedEvent event) {
    EventContext.setEventContext(
        event.getMetadata().getCorrelationId().toString(),
        event.getMetadata().getEventId().toString()
    );
    UpdateSecurityDetailsDto dto = new UpdateSecurityDetailsDto(null, event.getIsLockUser());
    securityDetailsService.updateSecurityCredentials(
        event.getUserId().toString().toUpperCase(), dto);
}

4. Port & Path Mapping Reference