Trigger.dev + LangSmith legacy Telemetry Demo

This project demonstrates how to integrate Trigger.dev with LangSmith using their legacy telemetry SDK for AI telemetry and observability. It includes a sample task that generates text using OpenAI's GPT-4o model with AI SDK 4 while capturing comprehensive telemetry data in LangSmith using the LangSmith SDK.

Features

• Trigger.dev Integration: Background task processing with retry logic and scaling
• LangSmith SDK Integration: Native LangSmith SDK for enhanced telemetry and tracing
• AI SDK 4: Modern AI SDK for OpenAI model interactions
• OpenAI Integration: Text generation using GPT-4o model
• OTEL Telemetry: OpenTelemetry-based observability with LangSmith
• TypeScript Support: Fully typed codebase for better development experience

Prerequisites

Before getting started, make sure you have:

1. Node.js (version 18 or higher)
2. npm or yarn package manager
3. Trigger.dev account - Sign up here
4. OpenAI API key - Get one here
5. LangSmith account - Sign up here

Setup Instructions

1. Clone and Install Dependencies

# Install dependencies
npm install

2. Environment Configuration

Create a .env file in the root directory with the following variables:

# Trigger.dev
TRIGGER_PROJECT_REF=your_trigger_project_ref_here

# OpenAI
OPENAI_API_KEY=your_openai_api_key_here

# LangSmith
LANGSMITH_API_KEY=your_langsmith_api_key_here
LANGSMITH_PROJECT=your_project_name_here

Getting Your Env vars:

Trigger.dev:

1. Go to your Trigger.dev dashboard
2. Navigate to your project
3. Copy the "Project Reference" (starts with proj_) from the project settings

OpenAI:

1. Visit OpenAI API Keys
2. Create a new secret key
3. Copy the key (starts with sk-)

LangSmith:

1. Go to LangSmith Settings
2. Create an API key in the "API Keys" section
3. Create a project or use an existing one

3. Development Setup

Start the development server:

npm run dev

This will:

• Start the Trigger.dev development environment
• Watch for file changes and reload automatically
• Connect to your Trigger.dev project
• Enable local task execution

4. Testing the Task

Once the development server is running, you can trigger the task:

1. Via Trigger.dev Dashboard:

   • Go to your project dashboard
   • Find the "test-telemetry" task
   • Click "Test" and provide a payload like: {"input": "test message"}

2. Via API (using curl):

   curl -X POST "https://api.trigger.dev/api/v1/tasks/test-telemetry/trigger" \
     -H "Authorization: Bearer YOUR_TRIGGER_SECRET_KEY" \
     -H "Content-Type: application/json" \
     -d '{"input": "Hello from API"}'

Project Structure

├── src/
│   └── trigger/
│       └── task.ts          # Main task definition
├── trigger.config.ts        # Trigger.dev configuration
├── package.json            # Dependencies and scripts
├── tsconfig.json           # TypeScript configuration
├── .env                    # Environment variables (create this)
└── README.md              # This file

How It Works

The Task (src/trigger/task.ts)

The main task does the following:

1. Receives Input: Takes a string input parameter
2. Creates Trace: Uses Trigger.dev's logger to create a trace span
3. Calls OpenAI: Uses AI SDK 4 to generate text with GPT-4o
4. Captures Telemetry: Automatically sends trace data to LangSmith via OTEL
5. Returns Results: Logs the generated text

Telemetry Configuration

The trigger.config.ts file configures:

• LangSmith OTEL Integration: Uses LangSmith SDK's initializeOTEL() for seamless telemetry setup
• Automatic Span Export: Configures the default LangSmith span exporter
• Retry Logic: Handles failures with exponential backoff
• Process Management: Optimizes performance with keep-alive settings
• Machine Scaling: Uses small-2x machines for processing

Monitoring and Observability

LangSmith Dashboard

After running tasks, you can view detailed telemetry in your LangSmith dashboard:

1. Traces: See complete execution traces with timing
2. Model Calls: Monitor OpenAI API usage and performance
3. Errors: Debug failed requests with full context
4. Metrics: Track usage patterns and costs

Trigger.dev Dashboard

Monitor your tasks in the Trigger.dev dashboard:

1. Task Runs: View execution history and status
2. Logs: Access detailed execution logs
3. Performance: Monitor execution times and resource usage
4. Scheduling: Set up recurring or delayed task execution

Deployment

To deploy your tasks to production:

npm run deploy

This will:

• Build your TypeScript code
• Deploy to Trigger.dev's infrastructure
• Make your tasks available for production use

Environment Variables Reference

Variable	Description	Required
TRIGGER_PROJECT_REF	Trigger.dev project reference ID	✅
OPENAI_API_KEY	OpenAI API key for GPT model access	✅
LANGSMITH_API_KEY	LangSmith API key for telemetry	✅
LANGSMITH_PROJECT	LangSmith project name	✅

Troubleshooting

Common Issues

1. "Project not found" error

   • Verify your TRIGGER_PROJECT_REF is correct
   • Check that the project reference in your .env file matches your Trigger.dev project

2. OpenAI API errors

   • Ensure your OPENAI_API_KEY is valid and has sufficient credits
   • Check that you have access to the GPT-4o model

3. LangSmith telemetry not appearing

   • Verify your LANGSMITH_API_KEY and LANGSMITH_PROJECT are correct
   • Check that your LangSmith project exists

4. Development server won't start

   • Make sure you're using Node.js 18 or higher
   • Try deleting node_modules and running npm install again

5. Telemetry configuration issues

   • Ensure the LangSmith SDK is properly initialized in trigger.config.ts
   • Verify that the OTEL setup is not conflicting with other telemetry tools

Getting Help

• Trigger.dev: Documentation | Discord
• LangSmith: Documentation | Support
• OpenAI: Documentation | Support

Key Technologies

AI SDK 4

This project uses AI SDK version 4.3.19, which provides:

• Improved OpenAI integration
• Better TypeScript support
• Enhanced streaming capabilities
• Optimized performance

LangSmith SDK

The LangSmith SDK (v0.3.67) enables:

• Native OTEL integration via initializeOTEL()
• Automatic span collection and export
• Seamless telemetry configuration
• Enhanced debugging capabilities

Next Steps

• Explore more complex AI workflows
• Set up scheduled tasks for regular processing
• Add error handling and retry logic
• Implement webhook triggers for real-time processing
• Scale to multiple tasks and workflows
• Experiment with different AI models and providers
• Add custom telemetry and metrics