AWS Bedrock Integration Guide¶

Overview¶

This guide provides detailed instructions for integrating and deploying models to AWS Bedrock using IdeaWeaver. AWS Bedrock is a fully managed service that makes it easy to build and scale generative AI applications with foundation models.

Table of Contents¶

Prerequisites
Setup
Model Requirements
Deployment Process
Troubleshooting
Best Practices

Prerequisites¶

AWS Account Setup¶

AWS Account with Bedrock access enabled
IAM Role with necessary permissions
S3 Bucket in us-east-1 region

Required Permissions¶

IAM Role Policy¶

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "bedrock:*"
            ],
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": [
                "s3:GetObject",
                "s3:ListBucket"
            ],
            "Resource": [
                "arn:aws:s3:::aws-bedrock-ideaweaver-bucket-use1",
                "arn:aws:s3:::aws-bedrock-ideaweaver-bucket-use1/*"
            ]
        }
    ]
}

S3 Bucket Policy¶

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Sid": "AllowBedrockAccess",
            "Effect": "Allow",
            "Principal": {
                "Service": "bedrock.amazonaws.com"
            },
            "Action": [
                "s3:GetObject",
                "s3:ListBucket"
            ],
            "Resource": [
                "arn:aws:s3:::aws-bedrock-ideaweaver-bucket-use1",
                "arn:aws:s3:::aws-bedrock-ideaweaver-bucket-use1/*"
            ]
        }
    ]
}

Setup¶

Create S3 Bucket

aws s3api create-bucket \
  --bucket aws-bedrock-ideaweaver-bucket-use1 \
  --region us-east-1

Configure Bucket Settings

# Disable ACLs
aws s3api put-bucket-ownership-controls \
  --bucket aws-bedrock-ideaweaver-bucket-use1 \
  --ownership-controls Rules=[{ObjectOwnership=BucketOwnerEnforced}]

Create IAM Role

aws iam create-role \
  --role-name bedrock-admin-role \
  --assume-role-policy-document file://trust-policy.json

Model Requirements¶

Supported Architectures¶

llama
mistral
t5
mixtral
gpt_bigcode
mllama
qwen2_vl
qwen2
qwen2_5_vl

Model Structure¶

my-qwen2-model/
├── config.json
├── pytorch_model.bin
├── tokenizer.json
└── tokenizer_config.json

Deployment Process¶

1. Prepare Model¶

# Download supported model
git lfs install
git clone https://huggingface.co/Qwen/Qwen2-0.5B my-qwen2-model

2. Upload to S3¶

# Upload model files
aws s3 sync ./my-qwen2-model s3://aws-bedrock-ideaweaver-bucket-use1/bedrock-models/my-qwen2-model/

3. Deploy to Bedrock¶

# Deploy using IdeaWeaver
ideaweaver deploy bedrock \
  --model-path ./my-qwen2-model \
  --model-name my-qwen2-model \
  --s3-bucket aws-bedrock-ideaweaver-bucket-use1 \
  --region us-east-1

Troubleshooting¶

Common Issues¶

Region Mismatch
Error: "Amazon Bedrock does not have access to the S3 location"
Solution: Ensure S3 bucket is in us-east-1
Permission Issues
Error: "Access Denied"
Solution: Verify IAM role and bucket policies
Model Architecture
Error: "Amazon bedrock does not support the architecture"
Solution: Use only supported architectures
Import Job Quotas
Error: "Your account does not have the quota limits"
Solution: Request quota increase

Debugging Steps¶

Verify S3 Access

aws s3 ls s3://aws-bedrock-ideaweaver-bucket-use1/bedrock-models/my-qwen2-model/

Check IAM Role

aws iam get-role --role-name bedrock-admin-role

Validate Model

ideaweaver bedrock validate-model ./my-qwen2-model

Best Practices¶

Model Selection
Start with smaller models (e.g., Qwen2-0.5B)
Verify architecture compatibility
Test locally before deployment
S3 Organization
Use consistent naming conventions
Organize models in subdirectories
Enable versioning for important models
Security
Use IAM roles with minimal permissions
Disable ACLs on S3 buckets
Regularly audit access policies
Monitoring
Track import job status
Monitor model performance
Set up CloudWatch alarms