# AWS Telemetry Demo
This sample demonstrates how to export telemetry (traces) from Genkit to AWS X-Ray
using the AWS plugin.
## Prerequisites
1. **AWS Credentials**: Configure via environment variables, credentials file, or IAM role
2. **AWS Region**: Set `AWS_REGION` environment variable
3. **IAM Permissions**: Attach `AWSXrayWriteOnlyPolicy` to your role/user
4. **Gemini API Key**: Set `GEMINI_API_KEY` (or use any other model provider)
## Quick Start
```bash
# Set required environment variables
export AWS_REGION=us-west-2
export GEMINI_API_KEY=your-api-key
# Run the demo
./run.sh
```
## Viewing Traces
1. Open the Dev UI (typically at http://localhost:4000)
2. Run the `hello_world` or `multi_step_demo` flow
3. View traces in AWS X-Ray Console:
https://console.aws.amazon.com/xray/home
Traces typically appear within 1-2 minutes.
## Flows
### `hello_world`
A simple flow that generates text. Demonstrates basic tracing.
### `multi_step_demo`
A multi-step flow with nested model calls. Shows how nested operations
appear as child spans in X-Ray.
## Environment Variables
| Variable | Required | Description |
|----------|----------|-------------|
| `AWS_REGION` | Yes | AWS region for X-Ray endpoint |
| `AWS_ACCESS_KEY_ID` | No* | AWS access key |
| `AWS_SECRET_ACCESS_KEY` | No* | AWS secret key |
| `AWS_PROFILE` | No* | AWS profile from credentials file |
| `GEMINI_API_KEY` | Yes | Gemini API key |
*At least one form of AWS credentials is required.
## IAM Policy
Ensure your IAM role/user has the `AWSXrayWriteOnlyPolicy` managed policy, or:
```json
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"xray:PutTraceSegments",
"xray:PutTelemetryRecords"
],
"Resource": "*"
}
]
}
```
## Documentation
- [AWS X-Ray](https://docs.aws.amazon.com/xray/)
- [Genkit Observability](https://genkit.dev/docs/observability)
