Use when building cloud-native apps. Keywords: kubernetes, k8s, docker, container, grpc, tonic, microservice, service mesh, observability, tracing, metrics, health check, cloud, deployment, 云原生, 微服务, 容器
Layer 3: Domain Constraints
| Domain Rule | Design Constraint | Rust Implication |
|---|---|---|
| 12-Factor | Config from env | Environment-based config |
| Observability | Metrics + traces | tracing + opentelemetry |
| Health checks | Liveness/readiness | Dedicated endpoints |
| Graceful shutdown | Clean termination | Signal handling |
| Horizontal scale | Stateless design | No local state |
| Container-friendly | Small binaries | Release optimization |
RULE: No local persistent state
WHY: Pods can be killed/rescheduled anytime
RUST: External state (Redis, DB), no static mut
RULE: Handle SIGTERM, drain connections
WHY: Zero-downtime deployments
RUST: tokio::signal + graceful shutdown
RULE: Every request must be traceable
WHY: Debugging distributed systems
RUST: tracing spans, opentelemetry export
From constraints to design (Layer 2):
"Need distributed tracing"
↓ m12-lifecycle: Span lifecycle
↓ tracing + opentelemetry
"Need graceful shutdown"
↓ m07-concurrency: Signal handling
↓ m12-lifecycle: Connection draining
"Need health checks"
↓ domain-web: HTTP endpoints
↓ m06-error-handling: Health status
| Purpose | Crate |
|---|---|
| gRPC | tonic |
| Kubernetes | kube, kube-runtime |
| Docker | bollard |
| Tracing | tracing, opentelemetry |
| Metrics | prometheus, metrics |
| Config | config, figment |
| Health | HTTP endpoints |
| Pattern | Purpose | Implementation |
|---|---|---|
| gRPC services | Service mesh | tonic + tower |
| K8s operators | Custom resources | kube-runtime Controller |
| Observability | Debugging | tracing + OTEL |
| Health checks | Orchestration | /health, /ready |
| Config | 1... |