Skip to content

CHANGELOG.md

Latest commit

 

History

History
78 lines (59 loc) · 3.12 KB

CHANGELOG.md

File metadata and controls

78 lines (59 loc) · 3.12 KB
Raw

Changelog

All notable changes to Monocle TypeScript will be documented in this file.

0.1.2 (2025-04-02)

Features

  • AWS Services
    • Added instrumentation for AWS Bedrock #40
    • Added instrumentation for AWS SageMaker #40
    • Added instrumentation for AWS OpenSearch #40

Bug Fixes

  • Fixed Azure Blob file naming issues #44
  • Fixed Windows-Linux compatibility issues #44

Maintenance

  • Removed unused files #43

0.1.0 (2025-03-25)

Features

  • Core Instrumentation

    • Added OpenTelemetry-based instrumentation for AI/ML frameworks
    • Implemented support for LangChain, LlamaIndex, and OpenAI
    • Created flexible metamodel for capturing AI operations and attributes
    • Added span handlers for workflow and operation tracking
    • Implemented context propagation for distributed tracing

  • AI Framework Support

    • LangChain

      • Implemented instrumentation for BaseChatModel, RunnableParallel, RunnableSequence
      • Added support for VectorStoreRetriever and PromptTemplate operations
      • Created inference and retrieval output processors

    • LlamaIndex

      • Implemented instrumentation for VectorIndexRetriever, RetrieverQueryEngine
      • Added support for OpenAI and BaseLLM methods
      • Created dedicated extraction methods for LlamaIndex specific objects

    • OpenAI

      • Added direct instrumentation for OpenAI client operations
      • Implemented input message extraction and response processing

  • Exporters

    • Added multiple exporter options for telemetry data:
      • Console exporter for development and debugging
      • File exporter for local analysis
      • AWS S3 exporter for cloud storage
      • Azure Blob Storage exporter for Microsoft environments
      • Okahu exporter for centralized analysis

  • Utilities

    • Implemented metadata extraction for token usage tracking
    • Added utilities for vector store deployment identification
    • Created helper methods for message extraction from various AI frameworks
    • Added support for identifying infrastructure environment (AWS Lambda, Azure, etc.)

  • Lambda Support

    • Implemented special handling for AWS Lambda environments
    • Added Lambda extension for asynchronous processing
    • Created specialized task processor for efficient telemetry export

  • Developer Experience

    • Added comprehensive documentation with examples
    • Created sample implementations and output processors
    • Added debug logging with MONOCLE_DEBUG environment variable control
    • Implemented custom instrumentation examples for Gemini and OpenAI

Documentation

  • Added custom instrumentation guide with examples
  • Created examples for vector database integration
  • Added sample output processors for different AI frameworks
  • Provided sample traces showing the correlation between processors and telemetry