簡介
本文透過 3 個橫切關注點完善了微服務平台:Redis 快取減少資料庫負載,用於 Kubernetes 就緒/活躍性探針的運行狀況檢查,以及 API 網關作為前端的單一入口點。
Redis 緩存
依賴關係
<dependency>
<groupId>io.quarkus</groupId>
<artifactId>quarkus-cache</artifactId>
</dependency>
<dependency>
<groupId>io.quarkus</groupId>
<artifactId>quarkus-redis-cache</artifactId>
</dependency>
配置
# Dev Services tự động start Redis container
# Production:
%prod.quarkus.redis.hosts=redis://localhost:6379
%prod.quarkus.redis.password=${REDIS_PASSWORD}
# Cache TTL
quarkus.cache.redis.product-cache.ttl=5M
quarkus.cache.redis.category-cache.ttl=30M
quarkus.cache.redis.product-list-cache.ttl=2M
@CacheResult — 快取方法返回
import io.quarkus.cache.CacheResult;
import io.quarkus.cache.CacheInvalidate;
import io.quarkus.cache.CacheInvalidateAll;
import io.quarkus.cache.CacheKey;
@ApplicationScoped
public class ProductService {
// Cache kết quả findById
@CacheResult(cacheName = "product-cache")
public ProductDTO getById(@CacheKey Long id) {
Log.infof("Cache MISS: product %d", id);
Product product = productRepo.findByIdOptional(id)
.orElseThrow(() ->
new ResourceNotFoundException("Product", id));
return ProductDTO.from(product);
}
// Invalidate khi update
@CacheInvalidate(cacheName = "product-cache")
@Transactional
public ProductDTO update(@CacheKey Long id,
UpdateProductRequest req) {
Product product = productRepo.findByIdOptional(id)
.orElseThrow(() ->
new ResourceNotFoundException("Product", id));
// ... update fields ...
return ProductDTO.from(product);
}
// Invalidate khi delete
@CacheInvalidate(cacheName = "product-cache")
@Transactional
public void delete(@CacheKey Long id) {
productRepo.deleteById(id);
}
// Cache list với composite key
@CacheResult(cacheName = "product-list-cache")
public List<ProductListDTO> listByCategory(
@CacheKey String category,
@CacheKey int page,
@CacheKey int size) {
return productRepo
.findActive(category, null,
Sort.by("createdAt").descending())
.page(Page.of(page, size))
.list().stream()
.map(ProductListDTO::from).toList();
}
// Invalidate toàn bộ cache
@CacheInvalidateAll(cacheName = "product-list-cache")
@Transactional
public ProductDTO create(CreateProductRequest req) {
// ... create product ...
}
}
程式設計緩存
import io.quarkus.cache.Cache;
import io.quarkus.cache.CacheName;
import io.quarkus.cache.CaffeineCache;
@ApplicationScoped
public class CacheManager {
@Inject
@CacheName("product-cache")
Cache productCache;
public void warmUp() {
// Pre-load popular products
List<Product> popular = Product
.find("status = 'ACTIVE'",
Sort.by("viewCount").descending())
.page(Page.of(0, 100))
.list();
for (Product p : popular) {
productCache.as(CaffeineCache.class)
.put(p.id, CompletableFuture.completedFuture(
ProductDTO.from(p)));
}
}
public void evictAll() {
productCache.invalidateAll().await().indefinitely();
}
}
健康檢查
依賴關係(自動可用)
<dependency>
<groupId>io.quarkus</groupId>
<artifactId>quarkus-smallrye-health</artifactId>
</dependency>
自動端點
# Tất cả health checks
curl http://localhost:8081/q/health
# Liveness — service đang chạy?
curl http://localhost:8081/q/health/live
# Readiness — service sẵn sàng nhận traffic?
curl http://localhost:8081/q/health/ready
# Startup — service đã start xong?
curl http://localhost:8081/q/health/started
內建健康檢查
Quarkus 自動新增:
- 資料庫(資料來源):資料庫連線正常嗎?
- Kafka:經紀人已連線?
- Redis:快取已連線?
自訂健康檢查
import org.eclipse.microprofile.health.*;
@Liveness
@ApplicationScoped
public class ServiceLivenessCheck
implements HealthCheck {
@Override
public HealthCheckResponse call() {
return HealthCheckResponse
.named("Product Service")
.up()
.withData("version", "1.0.0")
.build();
}
}
@Readiness
@ApplicationScoped
public class DatabaseReadinessCheck
implements HealthCheck {
@Inject
ProductRepository productRepo;
@Override
public HealthCheckResponse call() {
try {
long count = productRepo.count();
return HealthCheckResponse
.named("Database")
.up()
.withData("products.count", count)
.build();
} catch (Exception e) {
return HealthCheckResponse
.named("Database")
.down()
.withData("error", e.getMessage())
.build();
}
}
}
@Readiness
@ApplicationScoped
public class ExternalServiceCheck
implements HealthCheck {
@Inject @RestClient
ProductServiceClient productClient;
@Override
public HealthCheckResponse call() {
try {
productClient.list(0, 1);
return HealthCheckResponse
.named("Product Service")
.up().build();
} catch (Exception e) {
return HealthCheckResponse
.named("Product Service")
.down()
.withData("error", e.getMessage())
.build();
}
}
}
健康檢查回應
{
"status": "UP",
"checks": [
{
"name": "Product Service",
"status": "UP",
"data": { "version": "1.0.0" }
},
{
"name": "Database",
"status": "UP",
"data": { "products.count": 150 }
},
{
"name": "Reactive PostgreSQL connections health check",
"status": "UP"
}
]
}
API 閘道 — Nginx / Envoy
Nginx 配置
# nginx.conf
upstream product-service {
server product-service:8080;
}
upstream order-service {
server order-service:8080;
}
upstream payment-service {
server payment-service:8080;
}
upstream keycloak {
server keycloak:8080;
}
server {
listen 80;
server_name api.ecommerce.xdev.asia;
# Rate limiting
limit_req_zone $binary_remote_addr
zone=api:10m rate=100r/s;
# Product Service
location /api/v1/products {
limit_req zone=api burst=20 nodelay;
proxy_pass http://product-service;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
}
# Order Service
location /api/v1/orders {
limit_req zone=api burst=10 nodelay;
proxy_pass http://order-service;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
}
# Payment Service
location /api/v1/payments {
limit_req zone=api burst=5 nodelay;
proxy_pass http://payment-service;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
}
# Keycloak
location /realms {
proxy_pass http://keycloak;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
}
# Health check aggregation
location /health {
access_log off;
default_type application/json;
return 200 '{"status":"UP","gateway":"nginx"}';
}
}
Docker 組合
services:
api-gateway:
image: nginx:1.25-alpine
ports:
- "80:80"
- "443:443"
volumes:
- ./nginx/nginx.conf:/etc/nginx/conf.d/default.conf
depends_on:
- product-service
- order-service
- payment-service
練習
1.為產品服務添加Redis快取(getById, listByCategory)
2.建立/更新/刪除產品時實現快取失效
3. 建立自訂健康檢查:資料庫、外部服務
4.設定Nginx API網關路由到服務
5.在Nginx中加入速率限制
6. 測試:啟動→就緒→活性探針
總結
- Redis 快取 (
@CacheResult,@CacheInvalidate) 減少資料庫負載 - 健康檢查 (
/q/health/live,/q/health/ready) 對於 Kubernetes 探針 - 自訂健康擴展
HealthCheck— 檢查資料庫、外部服務 - API 閘道 (Nginx) — 單一入口點、路由、速率限制、TLS 終止
- 速率限制保護後端服務免於濫用
下一篇文章:測試——@QuarkusTest 和 Testcontainers。