簡介
Quarkus 是從頭開始為 GraalVM Native Image 設計的。本機可執行檔在 ~20ms 內啟動（相對於 2-5s JVM），使用 ~30MB RSS （相對於 150-300MB JVM）。這是 Kubernetes 上微服務的一大優點：快速擴充、節省資源。
GraalVM 本機映像 — 運作原理
┌────────────────────────────┐
│ Java Application │
│ + All Dependencies │
│ + Quarkus Extensions │
└──────────┬─────────────────┘
│
▼
┌────────────────────────────┐
│ GraalVM Native Compiler │
│ (Ahead-of-Time) │
│ │
│ • Static Analysis │
│ • Dead Code Elimination │
│ • Heap Snapshotting │
│ • Substrate VM │
└──────────┬─────────────────┘
│
▼
┌────────────────────────────┐
│ Native Executable │
│ ~50MB standalone binary │
│ Startup: ~20ms │
│ RSS: ~30MB │
└────────────────────────────┘
建立原生鏡像
安裝 GraalVM
# SDKMAN (recommended)
sdk install java 21.0.6-graalce
sdk use java 21.0.6-graalce
# Verify
java -version
# openjdk version "21.0.6" 2025-01-21
# GraalVM CE 21.0.6+7.1
使用 Maven 構建
# Build native executable
./mvnw package -Dnative
# Output: target/product-service-1.0.0-runner
# (standalone binary, ~50MB)
# Run
./target/product-service-1.0.0-runner
# Quarkus started in 0.019s
在容器中建置（不需要本機 GraalVM）
# Dùng builder image — không cần cài GraalVM local
./mvnw package -Dnative \
-Dquarkus.native.container-build=true \
-Dquarkus.native.builder-image=\
quay.io/quarkus/ubi-quarkus-mandrel-builder-image:jdk-21
多階段 Dockerfile
原生映像檔 Dockerfile
# Stage 1: Build native executable
FROM quay.io/quarkus/ubi-quarkus-mandrel-builder-image:jdk-21 \
AS build
COPY --chown=quarkus:quarkus mvnw /code/mvnw
COPY --chown=quarkus:quarkus .mvn /code/.mvn
COPY --chown=quarkus:quarkus pom.xml /code/
COPY --chown=quarkus:quarkus src /code/src
USER quarkus
WORKDIR /code
RUN ./mvnw package -Dnative \
-DskipTests \
-Dquarkus.native.additional-build-args=\
--initialize-at-build-time
# Stage 2: Runtime image
FROM quay.io/quarkus/quarkus-micro-image:2.0
WORKDIR /work/
COPY --from=build /code/target/*-runner /work/application
RUN chmod 775 /work /work/application \
&& chown -R 1001 /work \
&& chmod -R "g+rwX" /work
EXPOSE 8080
USER 1001
ENTRYPOINT ["./application", "-Dquarkus.http.host=0.0.0.0"]
JVM Dockerfile（替代）
FROM eclipse-temurin:21-jre-alpine
ENV LANGUAGE='en_US:en'
COPY --chown=185 target/quarkus-app/lib/ /deployments/lib/
COPY --chown=185 target/quarkus-app/*.jar /deployments/
COPY --chown=185 target/quarkus-app/app/ /deployments/app/
COPY --chown=185 target/quarkus-app/quarkus/ /deployments/quarkus/
EXPOSE 8080
USER 185
ENV JAVA_OPTS_APPEND="-Dquarkus.http.host=0.0.0.0 \
-Djava.util.logging.manager=org.jboss.logmanager.LogManager"
ENV JAVA_APP_JAR="/deployments/quarkus-run.jar"
ENTRYPOINT ["java", ${JAVA_OPTS_APPEND}, "-jar", \
"/deployments/quarkus-run.jar"]
比較 JVM 與 Native
| 指標 | JVM 模式 | 本機模式 |
|---|---|---|
| 啟動時間 | 2-5秒 | 0.015-0.05秒 |
| 記憶體 (RSS) | 150-300 MB | 30-80 MB |
| Docker 映像 | 〜300 MB | 〜80 MB |
| 建置時間 | 10-30 秒 | 3-10 分鐘 |
| 峰值吞吐量 | 更高（JIT） | 降低約 10-20% |
| 熱身 | 需要熱身 | 不需要 |
| 反思 | 自由 | 需要配置 |
什麼時候使用 Native？
- ✅ 無伺服器/功能（冷啟動很重要）
- ✅ 微服務定期擴展/縮減
- ✅ 高容器密度（節省 RAM）
- ✅ CLI 工具
- ❌ 長時間運行的服務需要最大吞吐量 → JVM 模式更好
反射和資源配置
// Quarkus tự phát hiện reflection qua extensions
// Nhưng nếu dùng dynamic reflection thủ công:
@RegisterForReflection
public class ProductDTO {
public Long id;
public String name;
public BigDecimal price;
}
// Đăng ký class từ library bên ngoài:
@RegisterForReflection(targets = {
com.external.lib.SomeClass.class,
com.external.lib.AnotherClass.class
})
public class ReflectionConfig {}
Native 的 application.properties
# Resources cần include trong native image
quarkus.native.resources.includes=\
db/migration/*.sql,\
META-INF/resources/**,\
templates/**
# Native image build settings
quarkus.native.additional-build-args=\
-H:+ReportExceptionStackTraces,\
--initialize-at-build-time
本機測試
// Integration Test chạy trên native binary
@QuarkusIntegrationTest
class ProductResourceIT {
@Test
void testListProductsNative() {
given()
.when().get("/api/v1/products")
.then()
.statusCode(200);
}
@Test
void testHealthCheck() {
given()
.when().get("/q/health")
.then()
.statusCode(200)
.body("status", equalTo("UP"));
}
}
# Chạy native integration tests
./mvnw verify -Dnative
# Quarkus sẽ:
# 1. Build native executable
# 2. Start native binary
# 3. Run *IT.java tests against it
# 4. Shut down
Docker Compose — 所有服務
# docker-compose.prod.yml
services:
product-service:
build:
context: ./product-service
dockerfile: src/main/docker/Dockerfile.native
ports:
- "8081:8080"
environment:
QUARKUS_DATASOURCE_JDBC_URL: >-
jdbc:postgresql://postgres:5432/productdb
QUARKUS_DATASOURCE_USERNAME: product
QUARKUS_DATASOURCE_PASSWORD: ${DB_PASSWORD}
QUARKUS_OIDC_AUTH_SERVER_URL: >-
http://keycloak:8080/realms/ecommerce
depends_on:
postgres:
condition: service_healthy
keycloak:
condition: service_healthy
order-service:
build:
context: ./order-service
dockerfile: src/main/docker/Dockerfile.native
ports:
- "8082:8080"
environment:
QUARKUS_DATASOURCE_JDBC_URL: >-
jdbc:postgresql://postgres:5432/orderdb
QUARKUS_DATASOURCE_USERNAME: order
QUARKUS_DATASOURCE_PASSWORD: ${DB_PASSWORD}
QUARKUS_REST_CLIENT_PRODUCT_SERVICE_URL: >-
http://product-service:8080
KAFKA_BOOTSTRAP_SERVERS: kafka:9092
depends_on:
postgres:
condition: service_healthy
kafka:
condition: service_healthy
payment-service:
build:
context: ./payment-service
dockerfile: src/main/docker/Dockerfile.native
ports:
- "8083:8080"
environment:
QUARKUS_DATASOURCE_JDBC_URL: >-
jdbc:postgresql://postgres:5432/paymentdb
KAFKA_BOOTSTRAP_SERVERS: kafka:9092
notification-service:
build:
context: ./notification-service
dockerfile: src/main/docker/Dockerfile.native
ports:
- "8084:8080"
environment:
KAFKA_BOOTSTRAP_SERVERS: kafka:9092
QUARKUS_MAILER_HOST: mailhog
QUARKUS_MAILER_PORT: 1025
postgres:
image: postgres:17
environment:
POSTGRES_MULTIPLE_DATABASES: >-
productdb,orderdb,paymentdb
POSTGRES_PASSWORD: ${DB_PASSWORD}
volumes:
- pgdata:/var/lib/postgresql/data
- ./init-multiple-dbs.sh:\
/docker-entrypoint-initdb.d/init.sh
healthcheck:
test: ["CMD-SHELL", "pg_isready -U postgres"]
interval: 5s
timeout: 5s
retries: 5
keycloak:
image: quay.io/keycloak/keycloak:26.2
command: start-dev --import-realm
ports:
- "8180:8080"
environment:
KC_DB: postgres
KC_DB_URL: jdbc:postgresql://postgres:5432/keycloak
KC_DB_USERNAME: keycloak
KC_DB_PASSWORD: ${DB_PASSWORD}
KEYCLOAK_ADMIN: admin
KEYCLOAK_ADMIN_PASSWORD: ${KEYCLOAK_ADMIN_PASSWORD}
volumes:
- ./keycloak/realms:/opt/keycloak/data/import
healthcheck:
test: ["CMD-SHELL",
"exec 3<>/dev/tcp/localhost/8080"]
interval: 10s
timeout: 5s
retries: 10
kafka:
image: confluentinc/cp-kafka:7.8.0
ports:
- "9092:9092"
environment:
KAFKA_NODE_ID: 1
KAFKA_PROCESS_ROLES: controller,broker
KAFKA_LISTENERS: >-
PLAINTEXT://0.0.0.0:9092,CONTROLLER://0.0.0.0:9093
KAFKA_CONTROLLER_QUORUM_VOTERS: 1@kafka:9093
KAFKA_CONTROLLER_LISTENER_NAMES: CONTROLLER
CLUSTER_ID: "MkU3OEVBNTcwNTJENDM2Qk"
healthcheck:
test: ["CMD", "kafka-broker-api-versions",
"--bootstrap-server", "localhost:9092"]
interval: 10s
timeout: 10s
retries: 5
nginx:
image: nginx:alpine
ports:
- "80:80"
volumes:
- ./nginx/nginx.conf:/etc/nginx/nginx.conf:ro
depends_on:
- product-service
- order-service
- payment-service
jaeger:
image: jaegertracing/jaeger:2
ports:
- "16686:16686" # UI
- "4317:4317" # OTLP gRPC
prometheus:
image: prom/prometheus:v3.2.1
ports:
- "9090:9090"
volumes:
- ./prometheus/prometheus.yml:\
/etc/prometheus/prometheus.yml:ro
grafana:
image: grafana/grafana:11.5.2
ports:
- "3000:3000"
volumes:
- ./grafana/dashboards:/var/lib/grafana/dashboards
volumes:
pgdata:
練習
- 為產品服務建置本機可執行檔：
./mvnw package -Dnative - 比較啟動時間與記憶體：JVM vs Native
- 為原生映像檔編寫多階段Dockerfile
- 跑步
@QuarkusIntegrationTest在本機二進位檔案上 - Docker Compose啟動整個系統（5個服務+基礎設施）
總結
- GraalVM Native Image：啟動 ~20ms，RSS ~30MB — 非常適合微服務
- 多階段 Dockerfile：建置階段 (Mandrel) → 執行階段 (quarkus-micro-image ~80MB)
@RegisterForReflection對於使用動態反射的類@QuarkusIntegrationTest在實際的本機二進位檔案上進行測試- Docker Compose 用於開發/登台，Kubernetes 用於生產（第 22 課）
下一篇文章：Kubernetes 部署和 CI/CD — 部署到 Kubernetes 叢集。