Google's Agent-to-Agent (A2A) protocol is a powerful framework that enables AI agents to communicate and collaborate effectively. This project demonstrates how to implement A2A with comprehensive observability using Grafana, showcasing a health club service agent as an example.
A2AJava is the Java implementation of the Google A2A protocol you can check the project here .

The system can provides comprehensive visibility into agent interactions:

• Task Lifecycle: Track tasks from initiation to completion
• Real-time Status: Monitor active tasks and their current states
• Execution Metrics: Measure task duration, success rates, and failure patterns
• Task Dependencies: Visualize task chains and dependencies between agents

• Input Parameters: Monitor all parameters sent to agent tasks
• Parameter Validation: Track parameter validation success/failure rates
• Parameter Usage Patterns: Analyze common parameter combinations
• Type Safety: Monitor parameter type mismatches and conversion issues

• Request/Response Tracking: Complete visibility of all JSON-RPC messages
• Payload Analysis: Monitor request/response payload sizes and patterns
• Error Rates: Track RPC failures and error types
• Latency Metrics: Measure round-trip times for RPC calls
• Protocol Compliance: Ensure adherence to A2A protocol standards

The A2A protocol facilitates:

• Standardized Communication: Enables AI agents to expose their capabilities and interact using a common protocol
• Real-time Interaction: Supports streaming capabilities for live agent-to-agent communication
• Capability Discovery: Agents can dynamically discover and utilize each other's functions
• Scalable Architecture: Supports multiple agents working together in a distributed system

In a multi-agent system, observability is crucial for:

• Performance Monitoring: Track each agent's response times and throughput
• Behavior Analysis: Understand how agents interact and identify patterns
• Problem Detection: Quickly identify issues in agent communication
• Resource Optimization: Monitor resource usage and optimize agent deployment

This implementation showcases a Spring Boot-based health club service agent with real-time metrics streaming to Grafana dashboards, demonstrating how to build observable, intelligent services using the A2A protocol.

• Health Club Service Management

  • Gym appointment booking
  • Squash court reservations
  • Swimming session scheduling
  • Fitness class registration
  • Appointment cancellation and rescheduling

• A2A Protocol Integration

  • Real-time agent communication
  • Streaming capabilities
  • Agent card exposure via REST endpoints

• Monitoring & Metrics

  • Prometheus metrics integration
  • Grafana dashboard visualization
  • Real-time performance monitoring
  • Method-level timing and tracking

• Java 23
• Spring Boot 3.2.4
• Micrometer (for metrics)
• Prometheus
• Grafana
• Google A2A Protocol
• WebSocket support
• Lombok

1. Clone the repository:

   git clone https://github.com/vishalmysore/a2ajava-grafana.git

2. Navigate to the project directory:

   cd a2ajava-grafana

3. Build the project with Maven:

   mvn clean install

4. Run the application:

   mvn spring-boot:run

The application uses the following properties in application.properties:

spring.application.name=grafana-test
management.endpoints.web.exposure.include=prometheus
management.endpoint.prometheus.enabled=true

The A2A agent is exposed through the AgentCardController at the well-known path. The service provides various health club-related actions that can be accessed through the A2A protocol.

1. Access the Agent Card

   • The agent card is available at the well-known path configured in AgentCardController
   • View the agent in the Google A2A client:

   

2. Trigger Health Club Actions

   • Book appointments using the A2A protocol
   • View confirmations in the interface:

   

3. Monitor Activity

   • Access Prometheus metrics at /actuator/prometheus:

   

   • View metrics in Grafana dashboards
   • Track real-time performance data

• RPC call counts
• Method duration timings
• Service performance metrics
• System health indicators

1. Access your Grafana dashboard

2. Import the provided dashboards:

   

   

3. Monitor your health club service metrics in real-time

Contributions are welcome! Please feel free to submit a Pull Request.