Skip to main contentWhat is Tracing?
Tracing in PayLink is a comprehensive system that records detailed information about every payment operation executed by your AI agents. Each trace captures the complete context of a payment operation, including inputs, outputs, timing, status, and metadata, giving you full visibility into your payment workflows.
Key Characteristics
- Project Association: Every trace belongs to exactly one project
- Optional API Key Association: Traces can be associated with an API key when created via API
- Comprehensive Context: Traces capture tool execution details, request context, environment info, and metadata
- Unique Trace ID: Each trace has a unique identifier for tracking and correlation
- Payment-Focused: Designed specifically to track payment provider operations and tool executions
Traces record comprehensive information about each payment operation:
- Tool Information: Name of the tool/operation being executed (e.g.,
stk_push)
- Input Arguments: All parameters passed to the payment tool
- Output Response: The result or response from the payment operation
- Execution Status: Success, error, or pending status
- Performance Metrics: Execution duration in milliseconds
- Payment Provider: Which payment provider was used (M-Pesa, Airtel Money, etc.)
- Request Context: Information about the request (method, path, headers, client IP)
- Environment Info: MCP protocol version, payment provider details
- Metadata: Additional custom information about the operation
How Tracing Works
Automatic Trace Creation
When your agent executes a payment operation:
- Tool Execution: Your agent calls a payment tool (e.g.,
stk_push)
- Trace Capture: PayLink automatically captures all relevant information
- Storage: The trace is stored and associated with your project
- Dashboard Access: You can view the trace in the PayLink dashboard
Trace Association
Traces are automatically associated with:
- Project: Every trace belongs to a specific project
- API Key (optional): Traces created via API can be associated with the API key used
- Organization (indirect): Through the project’s organization
Use Cases
Debugging Payment Issues
- Identify Failures: Quickly identify which payment operations failed and why
- Error Analysis: Review error messages and stack traces
- Parameter Validation: Check if incorrect parameters were passed to payment tools
- Response Times: Monitor how long payment operations take
- Bottleneck Identification: Identify slow payment operations
- Optimization: Use timing data to optimize your payment workflows
Audit and Compliance
- Transaction History: Maintain complete records of all payment operations
- Compliance Reporting: Generate reports for regulatory requirements
- Audit Trails: Track who executed what operations and when
Analytics and Insights
- Usage Patterns: Understand how your agents are using payment tools
- Success Rates: Track success and failure rates of payment operations
- Provider Performance: Compare performance across different payment providers
Viewing Traces
In the PayLink Dashboard
- Log in to your PayLink account
- Navigate to Project Tracing in the dashboard
- Select the project you want to review
- Browse and filter traces by date, status, tool name, or other criteria
- Click on a trace to view detailed information
Trace Details
Each trace view shows:
- Complete input arguments
- Full output response
- Execution status and duration
- Request context and environment information
- Payment provider details
- Timestamps for creation and updates
Trace Lifecycle
- Creation: Trace is created when a payment tool is executed
- Storage: Trace is stored and associated with the project
- Access: Traces are accessible through the dashboard or API
- Deletion: Traces are automatically deleted when their associated project is deleted
Important Notes
- Traces are automatically created—you don’t need to manually create them
- Traces can be created with either
project_id or project_name (when using API key authentication)
- When using
project_name, the system resolves it using the API key’s organization
- Traces are automatically deleted when their associated project is deleted
- API key association is optional but recommended for tracking and access control
- The
trace_id field is unique and can be used for external correlation