Microservices Architecture Rules

microservicesarchitecturedistributed-systemsbackend

Rules for designing, building and operating microservices systems

Целевые файлы

Файл	Формат
.cursorrules	plaintext
CLAUDE.md	markdown

Содержимое

Each service must own its own database; never share a database between two services. Define service boundaries around business domains (Bounded Contexts), not technical layers. Communicate synchronously via HTTP/gRPC only when a real-time response is required. Use asynchronous messaging (events/queues) for operations that can tolerate latency. Publish versioned event schemas; never make breaking changes to an existing event version. Each service must expose a /health endpoint returning current status and dependencies. Use circuit breakers for all synchronous calls to downstream services. Propagate trace IDs (W3C Trace Context) across all service boundaries for observability. Store service configuration in environment variables; use a secrets manager for credentials. Deploy each service independently; avoid deployment coupling between services. Write service-level integration tests that use real dependencies, not mocks. Define and document the SLA for each service (latency p99, availability target). Implement idempotent consumers for all message queue handlers to handle redelivery. Use an API gateway for all external traffic; never expose individual services directly.

CLI

npx mindaxis apply microservices-rules --target cursor --scope project

Используется в паках

Event-Driven Design

Scalable async event-driven systems

Открыть пак →

← Правила