TReNDS Automates Root-Cause Analysis with Amazon Bedrock

The TReNDS Center at Georgia State University has implemented an automated root-cause analysis system using Amazon Bedrock, reducing the time it takes to investigate errors from 15-30 minutes to near real-time.

The center's infrastructure is built on Amazon Web Services (AWS) and includes applications running on Amazon Elastic Kubernetes Service (Amazon EKS) with logs sent to Amazon CloudWatch. A subscription filter watches for error-level patterns and invokes a Lambda function, which runs a Strands Agent powered by Amazon Bedrock.

The agent uses the foundation model (FM) in Amazon Bedrock to investigate errors, fetch additional log context from the same container, retrieve relevant source code from GitHub, and reason through the root cause. The analysis is then published to an Amazon Simple Notification Service (Amazon SNS) topic for delivery to the team.

The system has been designed with data residency and compliance in mind, keeping log data and source code within the same AWS environment as the rest of the application. This is particularly important for TReNDS, which handles health-related research data that may fall under HIPAA requirements.