Can I use this skill for high-concurrency environments?

Yes. It uses fcntl.flock for thread-safe writes, ensuring that multiple processes can log data simultaneously to the same JSON file without risk of corruption.

Where is the telemetry data stored?

By default, telemetry data is stored in a structured JSON array at /registry/telemetry.json within the Betty Framework directory, though this path is configurable via environment variables.

Is there a performance overhead when using the decorator?

The overhead is minimal (typically less than 1ms per call), as it uses standard Python libraries and local file I/O without making external network calls.

How does telemetry.capture handle sensitive information?

The skill includes an input sanitization feature through the 'sanitize_keys' parameter, allowing developers to redact sensitive data like passwords or API keys before they are written to the log.

Telemetry Capture & Observability

Name: Telemetry Capture & Observability
Author: epieczko

byepieczko

Analytics & Monitoring

Captures comprehensive execution metrics and performance data for Betty Framework components to provide deep system observability.

About

The telemetry.capture skill is a specialized tool designed to monitor and log the behavior of Betty Framework skills and agents. It provides developers with a structured way to track execution status, timing data, and contextual metadata using thread-safe JSON logging. Whether used as a Python decorator, a context manager, or via the CLI, it enables deep visibility into system performance, error tracking, and usage patterns. This skill is essential for maintaining an audit trail and debugging complex AI workflows by providing a centralized registry of all component activities.

Key Features

ISO 8601 timestamping with full timezone support and duration tracking
Decorator pattern (@capture_telemetry) for seamless automatic integration
Structured query interface to analyze historical execution and error data
Thread-safe JSON logging with fcntl file locking for data integrity
Built-in input sanitization to redact sensitive data and secrets
0 GitHub stars

Use Cases

Auditing execution patterns and caller chains across multi-step Betty workflows
Debugging failed skill executions using detailed error messages and captured context
Monitoring performance and execution duration for long-running AI agent tasks

About

Key Features

ISO 8601 timestamping with full timezone support and duration tracking
Decorator pattern (@capture_telemetry) for seamless automatic integration
Structured query interface to analyze historical execution and error data
Thread-safe JSON logging with fcntl file locking for data integrity
Built-in input sanitization to redact sensitive data and secrets
0 GitHub stars

Use Cases

Auditing execution patterns and caller chains across multi-step Betty workflows
Debugging failed skill executions using detailed error messages and captured context
Monitoring performance and execution duration for long-running AI agent tasks