Getting Started with s3lim

s3lim is designed to be highly modular, allowing you to choose the deployment method that best aligns with your team’s AWS permissions, security constraints, and existing infrastructure.

To help you get started, we offer the following deployment methods and execution guides:

  1. Autopilot Deployment: The quickest, fully automated setup. Best if you want s3lim to configure everything—including S3 Inventory reports on your source bucket—automatically.
  2. Standard Deployment: Best if you already have S3 Inventory reports configured and delivered to a bucket. Supports both automated Managed IAM and custom Bring-Your-Own-IAM (BYO-IAM) roles.
  3. Multi-Region Deployment (StackSets): Best if you want to deploy s3lim across multiple regions and accounts simultaneously using AWS CloudFormation StackSets.
  4. Execution Architecture: Detailed guide on s3lim’s serverless Step Functions Distributed Map pipeline, worker lifecycle, and concurrency tuning.
  5. MCP Gateway Setup: Guide on how to enable the Amazon Bedrock AgentCore MCP Gateway (by setting EnableMCPGateway to true during deployment) to connect AI assistants.

Choosing a Deployment Method

  • Choose Autopilot if you want to optimize a bucket quickly without configuring S3 Inventory manually.
  • Choose Standard if S3 Inventory is already running on your buckets. Use Managed IAM for automated permissions or BYO-IAM if your security team requires pre-audited IAM roles.
  • Choose Multi-Region (StackSets) if you have data distributed across multiple AWS regions or accounts and want to manage a centralized analytics pipeline.
  • Understanding runtime architecture & concurrency? Read the Execution Architecture Guide.
  • Want natural language cost audits from AI agents? Enable the MCP Gateway Setup on your stack.

NOTE: Always deploy s3lim in the same AWS Region as the S3 bucket with the inventory file to avoid cross-region data transfer fees and potential timeout issues.