You need to enable JavaScript to run this app.
Observability Platform

Observability Platform

  • Documentation
    • Observability PlatformPractice and tutorialsBest practices of AI application monitoringReporting observability data of Eino-Assistant application
Copy page
Download PDF
Best practices of AI application monitoring
Reporting observability data of Eino-Assistant application
Copy page
Download PDF
Reporting observability data of Eino-Assistant application
Application Performance Management (APM) has been adapted to multiple CloudWeGo frameworks, enabling tracing for CloudWeGo-related services. This topic describes how to view the observability data of the Eino-Assistant application, with its observability data reporting to APM as an example.​
Prerequisites​
For the code repository and documentation of Eino-Assistant, refer to github_eino_assistant.​
Procedure​
Step 1: Deploy Eino-Assistant​
  • Method 1: Local deployment​
  1. In .env, configure the APM authentication information APMPLUS_APP_KEY. For information on how to obtain the AppKey, refer to How to obtain the AppKey for APM authentication?.​
  1. In .env, configure the endpoint and API Key for the large model.​
  1. Deploy Redis locally by executing the docker-compose up -d command.​
  1. Start the application by executing the go run cmd/einoagent/main.go command.​
  • Method 2: Deployment on Volcano Engine​
  1. Log in to the ModelArk console and create two models: one for Embedding and one for Chat.​
  1. Log in to the Database console, create a Redis Cluster, and obtain the private network address.​
  1. Log in to the VKE console, deploy the Dockerfile application, and configure the Server-side Monitoring endpoint, ModelArk endpoint, and Redis parameters.​
​
​
​
  1. In VKE, create a public Service of the LoadBalancer type and obtain its public access address.​
  1. Access the Eino-Assistant application via a browser and interact with it.​
Step 2: View observability data in APM​
  • View metrics​
  1. Log in to the APM console, and select the project and region in the top navigation bar.​
  1. In the left navigation bar, select Service Monitoring > Service List to go to the service list page.​
  1. Click the Eino-Assistant service name to go to the service management page, where you can view the metrics of Eino-Assistant. The metrics include large model metrics, golden signals, and runtime monitoring metrics.​
  • Large model metrics​
  • Golden signals​
  • Runtime monitoring metrics​
  • View traces​
  1. Click Trace Analysis in the left navigation bar, search for traces related to Eino-Assistant by service name, and view the trace list.​
  1. Click the target trace ID to view the details of a single trace. Three display modes are supported: flame graph, list, and topology map.​
Last updated: 2026.07.09 20:16:23