Skip to content

20240924, one-time fork from gabblegrid to create gabblegrid-mini which will be pared down to the playground, models and admin tabs

License

Notifications You must be signed in to change notification settings

UC-Berkeley-I-School/gabblegrid-mini

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

57 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GabbleGrid

GabbleGrid is a cutting-edge platform designed to enhance service resiliency by leveraging autonomous AI Agents. Our mission is to reduce the impact of service outages in complex IT infrastructures by providing a robust system that can predict and manage potential disruptions.

Key Features

  • Autonomous AI Agents: Deploy teams of AI agents that work collaboratively to monitor, predict, and act on potential service disruptions.
  • Service Outage Management: Reduce unplanned downtime caused by human error, technical issues, or cyberattacks through predictive analysis and automation.
  • Model Selection & Evaluation Platform: An easy-to-use platform for selecting and evaluating machine learning models tailored to your specific needs.
  • Real-time Log Analysis: Process and analyze unstructured log data using advanced Transformer-based models to detect anomalies.

Architecture and System Components

GabbleGrid's architecture is built on the following components:

  • Model Inference Engine: Runs real-time inference on log data to generate alerts and predictions.
  • Agent Teams: Specialized AI agents designed for tasks such as log retrieval, user interaction, and automated email generation.
  • Core Infrastructure: Hosted on EC2 instances with data stored on EBS, ensuring a scalable and robust environment.

How It Works

  1. Data Observation: GabbleGrid observes log data over specified time windows and predicts potential service disruptions based on learned patterns.
  2. Model Inference: Utilizing a Transformer-based model, the platform classifies logs as either normal or anomalous.
  3. Alert Generation: When anomalies are detected, the system generates actionable alerts, reducing the occurrence of false positives.
  4. Automation: AI agents work autonomously to handle remediation, reducing the need for manual intervention.

Why GabbleGrid?

  • High Precision: Focuses on reducing false positives to ensure that alerts are actionable and trustworthy.
  • Scalable: Designed to expand beyond initial use cases, including integration with other systems and the addition of new features.
  • State-of-the-Art Technology: Utilizes the latest advancements in AI, particularly Transformer models, to handle complex log data and long-range dependencies.

Future Roadmap

  • Enhanced Model Robustness: Continued focus on improving model precision and recall.
  • Broadened Scope: Expansion of GabbleGrid to support additional IT systems beyond the initial BG/L scope.
  • Feature Expansion: Introducing more features and visualizations to enhance user experience.

Installation

  1. Clone the repository:
    git clone https://github.com/gaurav8936/gabblegrid.git

About

20240924, one-time fork from gabblegrid to create gabblegrid-mini which will be pared down to the playground, models and admin tabs

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published