Skip to content

Dubbo service Apollo integration pattern

The Dubbo service Apollo integration pattern describes the architectural approach and workflow for decoupling configuration management from Dubbo microservice containers by integrating the Apollo configuration center. This pattern enables centralized configuration management, supports multi-environment deployments (Test/Production) using a single image, and facilitates dynamic configuration updates without requiring container rebuilds^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

Background and Motivation

Prior to integrating a configuration center like Apollo, Dubbo services often suffer from scattered configuration formats (XML, properties, etc.) and hard-coded settings within container images^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md]. This approach creates significant operational friction:

  • Redundant Packaging: Different environments (e.g., test, prod) require different configuration files, leading to the need to build and maintain separate images for each environment^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  • High Operational Cost: Updating a configuration requires rebuilding the image and redeploying containers, which is inefficient and error-prone^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  • Risk of Accidents: There is a high risk of mixing test environment configurations with production environment configurations, potentially causing production incidents^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

The core goal of this pattern is to ensure that "one image can be deployed to multiple environments", with configurations managed and injected externally via Apollo^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

Architecture

The integration typically follows the standard Apollo architecture, where the Dubbo service acts as a Client^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

Components

  • Apollo Portal: The management interface for configuring applications and namespaces^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  • Apollo ConfigService: Provides configuration reading interfaces to clients (Dubbo services)^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  • Apollo AdminService: Manages configuration modifications and pushes updates to the ConfigService^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  • ConfigDB: The database backend storing configuration data^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  • Dubbo Service (Client): Connects to the ConfigService to pull configurations. It also creates a local cached configuration file based on the remote settings^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

Data Flow

  1. Pull/Push: The Dubbo client maintains a long polling connection with the ConfigService to receive real-time configuration updates^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  2. Injection: When the Dubbo service starts, the Apollo client initializes, connects to the specified ConfigService, retrieves the relevant configuration (e.g., Zookeeper addresses, ports), and generates a local properties file (e.g., dubbo.properties) which the application then reads^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

Implementation Strategy

Environment Isolation

To achieve the goal of a single image for multiple environments, distinct environments (e.g., test and prod) are segregated using Kubernetes Namespaces and corresponding Apollo configurations^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

  • Database Isolation: Separate Apollo databases (e.g., ApolloConfigTestDB, ApolloConfigProdDB) are created for each environment^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  • Service Deployment: Dedicated instances of Apollo ConfigService and AdminService are deployed into specific Kubernetes Namespaces (e.g., test, prod)^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

Client Configuration

The Dubbo application requires specific environment variables to connect to the correct Apollo environment^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

  • AppId: Uniquely identifies the application in Apollo^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  • Environment (-Denv): Specifies the target environment (e.g., fat for test, pro for production)^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  • Meta Server (-Dapollo.meta): Specifies the address of the ConfigService. This varies by environment^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
    • Example for Test: -Dapollo.meta=http://config-test.od.com
    • Example for Prod: -Dapollo.meta=http://apollo-configservice:8080 (using internal K8s service DNS)^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

Delivery Workflow

  1. Code Commit: Developers commit code to a specific branch (e.g., apollo branch)^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  2. Build: A CI/CD pipeline (e.g., Jenkins) builds the Docker image based on the code. At this stage, the environment is generic^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  3. Deploy Test:
    • Update the Kubernetes Deployment in the test namespace with the new image tag^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
    • The container starts with C_OPTS="-Denv=fat -Dapollo.meta=http://config-test.od.com".
    • The service connects to the Test environment ConfigService and loads test configurations^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  4. Deploy Production:
    • Once testing is verified, update the Kubernetes Deployment in the prod namespace with the same image tag^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
    • The container starts with C_OPTS="-Denv=pro -Dapollo.meta=http://apollo-configservice:8080".
    • The service connects to the Production ConfigService and loads production configurations^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

Dynamic Configuration Updates

A key advantage of this pattern is the ability to modify runtime behavior without redeployment. For example, if a Dubbo service needs to switch its Zookeeper registry address or change its listening port^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md]:

  1. Modify: An administrator updates the configuration key (e.g., dubbo.port) in the Apollo Portal^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  2. Publish: The change is published to the environment^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  3. Restart: The Dubbo service pods are restarted (or rolled)^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].
  4. Effect: Upon restart, the Apollo client fetches the new configuration (e.g., port 20881) and applies it^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md].

Sources

^[400-devops__06-Kubernetes__k8s-paas__06.在K8S中集成Apollo配置中心.md]