AI Discovery
AI resource discovery has been added, enabling identification of:
- AI Services – View connect AI services.
- AI Logs – Captures detailed records of AI interactions, tracking input/output messages, token usage, and latency.
- AI Prompts – Tracks specific inputs provided to AI models, storing metadata such as temperature settings and response formats.
- AI Models – Identifies AI models in use, displaying performance metrics, associated prompts, and detailed execution logs.
AI Integration
- Bedrock Logging Integration – Log showing AI usage in AWS Bedrock. integration to start logging AI requests and responses.
- OpenAI Discovery Integration – Enables discovery of OpenAI-based models. This integration will discover AI models by usage.
Integrations
- Rust Library Added
- Wiz Integration Enhancement – Added the ability to upload findings for a selected app.
Actions
- Headers Support Added to Actions – Configure custom headers for API actions, enabling better control over authentication and request formatting.
Logging API
- Added support for bedrock logging of AI usage
AI Log Processor
- Added service and code to process AI logs.
- Added tags for language usage in prompts.
Event System
- Added AI object deletion support
- AI Event support added:
PaaS Events:
- LLM Discovered
- LLM Deleted
- Prompt Messages Discovered
- Prompt Messages Removed
Software Events:
- LLM Discovered
- LLM Deleted
- Prompt Discovered
- Prompt Deleted
- Prompt Messages Discovered
- Prompt Messages Removed
Repo Analysis
- Added discovery of prompts and models in code for OpenAI and anthropic in golang and python.
- Ruby Sinatra Support
- History Tracking Added
Bug Fixes
- General bug fixes and improvements made throughout the FireTail platform.
