Catalog
concept#Architecture#Integration#Software Engineering

Service Choreography

A decentralized coordination model for distributed services where individual services control interaction logic locally and interactions occur without a central controller.

Service choreography is a decentralized model for coordinating distributed services, where each component enforces interaction rules locally.
Established
High

Classification

  • High
  • Technical
  • Architectural
  • Intermediate

Technical context

Message brokers (Kafka, RabbitMQ)Observability tools (Jaeger, Prometheus)Schema registry or AsyncAPI tooling

Principles & goals

Distribute responsibility for flow logic across servicesDefine clear, versionable interaction contractsEnsure robust observability and monitoring of event flows
Build
Domain, Team

Use cases & scenarios

Compromises

  • Inconsistencies with insufficient compensation strategy
  • Hidden dependencies and implicit sequence prerequisites
  • Excessive observability effort without proper tooling
  • Write idempotent event handlers for safe retries
  • Use clear, versioned event schemas and a registry
  • Implement distributed tracing for traceability

I/O & resources

  • Explicit event or API contracts
  • Message broker or event infrastructure
  • Monitoring, tracing and versioning tools
  • Decoupled, autonomous services with defined interaction contracts
  • Events as integration artifacts
  • Distributed error and compensation paths

Description

Service choreography is a decentralized model for coordinating distributed services, where each component enforces interaction rules locally. Unlike orchestration, there is no central controller and coordination and error handling responsibilities are distributed. It is particularly suited to loosely coupled, domain-specific microservice architectures.

✔

  • Reduced central dependency increases resilience of individual components
  • Improved team autonomy through local decision authority
  • Scalability of individual reaction paths without a central bottleneck

✖

  • Increased complexity in error handling and compensation
  • Harder end-to-end testing due to distributed flows
  • Requires strict event and contract discipline between teams

  • Throughput per event topic

    Measures the number of events processed per time interval for a topic.

  • End-to-end latency

    Time from emitting an event to final processing and acknowledgement.

  • Failure and compensation rate

    Share of transactions requiring compensation or manual intervention.

Event-driven order pipeline

An online shop distributes order, payment and shipping logic to autonomous services that are choreographed via events.

Microservice saga as choreography

A saga implementation distributes compensation steps across involved services that react to each other's events.

IoT integration with decentralized control

Edge components and cloud services coordinate actions via events without a central orchestrator.

1

Analyze domain boundaries and identify relevant events.

2

Define event and API contracts and versioning rules.

3

Introduce event infrastructure, observability and compensation conventions.

⚠️ Technical debt & bottlenecks

  • Unstructured events without registry increase integration costs
  • Missing trace propagation hinders debugging
  • Ad-hoc compatibility fixes instead of proper versioning
Event broker capacityConsistency mechanismsMonitoring and trace infrastructure
  • Using choreography for highly transactional workloads without compensation
  • Lack of versioning leads to widespread incompatibilities
  • Insufficient monitoring so errors are not detectable
  • Underestimating the observability effort
  • Implicit sequence assumptions between independent services
  • Introducing compensation logic too late
Distributed systems architectureEvent design and versioningMonitoring, tracing and error compensation
Loosely coupled componentsDomain boundaries and team autonomyObservability and traceability
  • • Requirement for compatible event formats
  • • Network latency and asynchronous delays
  • • Regulatory requirements for transaction safety