Arconia Tracing

How to use OpenInference instrumentation for Spring AI with Arconia and export traces to Arize Phoenix.

Prerequisites

• Java 21 or higher
• (Optional) Phoenix API key if using Phoenix Cloud
• (Optional) Docker or Podman if using the Arconia Phoenix Dev Service

Add Dependencies

Gradle

Add the dependencies to your build.gradle:

dependencies {
    implementation 'io.arconia:arconia-openinference-ai-semantic-conventions'
    implementation 'io.arconia:arconia-opentelemetry-spring-boot-starter'

    implementation 'org.springframework.boot:spring-boot-starter-web'
    implementation 'org.springframework.ai:spring-ai-starter-model-mistral-ai'

    testAndDevelopmentOnly 'io.arconia:arconia-dev-services-phoenix'
}

Maven

Add the dependencies to your pom.xml:

<dependencies>
  <dependency>
      <groupId>io.arconia</groupId>
      <artifactId>arconia-openinference-ai-semantic-conventions</artifactId>
  </dependency>
  <dependency>
      <groupId>io.arconia</groupId>
      <artifactId>arconia-opentelemetry-spring-boot-starter</artifactId>
  </dependency>
  <dependency>
      <groupId>org.springframework.ai</groupId>
      <artifactId>spring-ai-starter-model-mistral-ai</artifactId>
  </dependency>
  <dependency>
      <groupId>org.springframework.boot</groupId>
      <artifactId>spring-boot-starter-webmvc</artifactId>
  </dependency>
  <dependency>
    <groupId>io.arconia</groupId>
    <artifactId>arconia-dev-services-phoenix</artifactId>
    <scope>runtime</scope>
    <optional>true</optional>
  </dependency>
</dependencies>

Setup Phoenix

Phoenix Dev Service

If you included the Arconia Phoenix Dev Service dependency as instructed in the previous step, your Spring Boot application will automatically provision a Phoenix service at startup time and connect to it. No extra code or configuration needed.

The application logs will show you the URL where you can access the Phoenix AI observability platform in your development environment.

...Phoenix UI: http://localhost:<port>

By default, traces are exported via OTLP using the HTTP/Protobuf format.

For more info on using Phoenix with Arconia, see Phoenix Dev Service.

Docker

Pull latest Phoenix image from Docker Hub:

docker pull arizephoenix/phoenix:latest

Run your containerized instance:

docker run -p 6006:6006 -p 4317:4317 arizephoenix/phoenix:latest

This command:

• Exposes port 6006 for the Phoenix web UI
• Exposes port 4317 for the OTLP gRPC endpoint (where traces are sent)

For more info on using Phoenix with Docker, see Docker.

Phoenix Cloud

Sign up for Phoenix:

1. Sign up for an Arize Phoenix account at https://app.phoenix.arize.com/login

2. Click Create Space, then follow the prompts to create and launch your space.

Set your Phoenix endpoint and API Key:

From your new Phoenix Space

1. Create your API key from the Settings page

2. Copy your Hostname from the Settings page

3. Set your endpoint and API key:

   export PHOENIX_API_KEY = "your-phoenix-api-key"
   export PHOENIX_COLLECTOR_ENDPOINT = "your-phoenix-endpoint"
   

Having trouble finding your endpoint? Check out Finding your Phoenix Endpoint

If you are using Phoenix Cloud or a self-hosted Phoenix, adjust the endpoint in the code as needed via the arconia.otel.exporter.otlp.endpoint=${PHOENIX_COLLECTOR_ENDPOINT} and arconia.otel.exporter.otlp.headers=Authorization=Bearer ${PHOENIX_API_KEY} properties. Alternatively, you can use the canonical OpenTelemetry Environment Variables: OTEL_EXPORTER_OTLP_ENDPOINT and OTEL_EXPORTER_OTLP_HEADERS.

Run Spring AI with Arconia

By instrumenting your application with Arconia, spans are automatically created whenever your AI models via Spring AI are invoked and sent to the Phoenix server for collection. Arconia plugs into Spring Boot and Spring AI without any code or configuration changes.

package io.arconia.demo;

import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.ai.chat.client.ChatClient;
import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RestController;

@SpringBootApplication
public class ArconiaTracingApplication {
    public static void main(String[] args) {
        SpringApplication.run(ArconiaTracingApplication.class, args);
    }
}

@RestController
class ChatController {

    private static final Logger logger = LoggerFactory.getLogger(ChatController.class);
    private final ChatClient chatClient;

    ChatController(ChatClient.Builder chatClientBuilder) {
        this.chatClient = chatClientBuilder.clone().build();
    }

    @GetMapping("/chat")
    String chat(String question) {
        logger.info("Received question: {}", question);
        return chatClient
                .prompt(question)
                .call()
                .content();
    }
}

Full example: https://github.com/arconia-io/arconia-examples/tree/main/arconia-openinference

Observe

Once configured, your OpenInference traces will be automatically sent to Phoenix where you can:

• Monitor Performance: Track latency, throughput, and error rates
• Analyze Usage: View token usage, model performance, and cost metrics
• Debug Issues: Trace request flows and identify bottlenecks
• Evaluate Quality: Run evaluations on your LLM outputs

Resources

Full Example

Arconia OpenInference Semantic Conventions (docs)