Skip to content

checklist.md

Latest commit

 

History

History
235 lines (187 loc) · 7.88 KB

checklist.md

File metadata and controls

235 lines (187 loc) · 7.88 KB

AWS NYC Workshop - Quick Checklist

Pre-Workshop Setup

Prerequisites Check

  • AWS Account with the following IAM managed policies attached (or equivalent permissions):
    • AmazonEC2FullAccess — VPC, subnets, security groups, EC2 instances
    • IAMFullAccess — IAM roles and instance profiles for EC2
    • AWSCloudFormationFullAccess — stack create/update/delete
    • CloudWatchLogsFullAccess — log groups for MemMachine
    • ElasticLoadBalancingFullAccess — Network Load Balancer
    • AmazonAPIGatewayAdministrator — HTTP API, VPC Link, routes
    • AmazonS3ReadOnlyAccess — access CFT template from S3 (if hosted in a bucket)
    • AmazonBedrockFullAccess — model invocation and listing
  • AWS CLI installed (aws --version)
  • AWS credentials configured (aws sts get-caller-identity)
  • EC2 Key Pair created in your region
  • Bedrock model access enabled
  • Python 3.9+ installed (Python 3.9 or higher)

Required Information

  • AWS Access Key ID
  • AWS Secret Access Key
  • EC2 Key Pair name
  • AWS Region (e.g., us-west-2)
  • Bedrock model IDs (embedding & language)

Part 1: Deploy MemMachine

Step 1: CloudFormation Deployment

  • Locate bedrock-cft.yml template (included in this directory)
  • Set environment variables or prepare parameters
  • Deploy stack via AWS CLI or Console
  • Wait for stack creation (5-10 minutes)
  • Retrieve stack outputs (ApplicationURL, PublicIP, Neo4j URL)

Step 2: Verification

  • Test health endpoint via API Gateway: curl https://YOUR_API_GATEWAY_URL/api/v2/health
  • Verify API docs accessible: https://YOUR_API_GATEWAY_URL/docs
  • Check Neo4j browser: http://PUBLIC_IP:7474

Note: Port 8080 on the EC2 instance is only accessible within the VPC. Use the API Gateway URL for all MemMachine API access.

Key Outputs to Save:

  • ApplicationURL (API Gateway): _____________
  • PublicIP: _____________
  • Neo4j URL: http://_____________:7474

Part 2: Build Chatbot

Step 1: Environment Setup

  • Navigate to workshop directory: cd aws_nyc
  • Create virtual environment (optional)
  • Install dependencies: pip install -r requirements.txt
    • Includes: boto3, requests, python-dotenv, streamlit

Step 2: Configuration

  • Copy .env.example to .env
  • Set MEMORY_SERVER_URL to the ApplicationURL (API Gateway URL) from Part 1
  • Set AWS credentials
  • Set Bedrock model ID
  • Set ORG_ID and PROJECT_ID

Step 3: Chatbot Code

  • Use without_memory.py (stateless, no memory - Streamlit web UI)
  • Use with_memory.py (with MemMachine memory - Streamlit web UI)
  • Test chatbot without memory: streamlit run without_memory.py
  • Test chatbot with memory: streamlit run with_memory.py
  • Compare the difference in the web interface!
  • Try switching models in the memory-enabled chatbot dropdown

Part 3: Compare Without vs With Memory

Test Chatbot Without Memory

  • Run: streamlit run without_memory.py
  • Web interface opens in browser
  • Send: "My name is [YourName]"
  • Ask: "What's my name?"
  • Notice: It doesn't remember! (This is expected)

Test Chatbot With Memory

  • Run: streamlit run with_memory.py
  • Web interface opens in browser
  • Send: "My name is [YourName]"
  • Send: "I love [something]"
  • Ask: "What's my name?"
  • Verify chatbot remembers!
  • Try switching models using the dropdown
  • Verify memory persists across model switches!

Memory Persistence Test

  • Restart chatbot
  • Ask: "What did we talk about earlier?"
  • Verify memory persists across sessions

Advanced Test

  • Have a multi-turn conversation
  • Ask follow-up questions
  • Verify context is maintained
  • Switch models mid-conversation using dropdown
  • Verify memory context is retained across different models
  • Toggle "Show Memory Context" to see retrieved memories

Part 4: Troubleshooting

Common Issues

  • Connection refused: Verify MEMORY_SERVER_URL is the API Gateway URL (not the EC2 IP directly)
  • Bedrock access denied: Verify model access in AWS Console
  • Memory not working: Check API responses, verify URLs
  • Stack creation failed: Check CloudFormation events

Verification Commands

# Test MemMachine via API Gateway
curl https://YOUR_API_GATEWAY_URL/api/v2/health

# Test MemMachine directly (only works from within the VPC / SSH into EC2)
# curl http://localhost:8080/api/v2/health

# Test Bedrock
aws bedrock list-foundation-models --region us-west-2

# Check stack status
aws cloudformation describe-stacks --stack-name YOUR_STACK_NAME

# Get stack outputs (including ApplicationURL)
aws cloudformation describe-stacks --stack-name YOUR_STACK_NAME --query 'Stacks[0].Outputs'

Workshop Goals Checklist

By the end of the workshop, you should have:

  • MemMachine deployed on AWS EC2 (with API Gateway)
  • Working chatbot using Bedrock
  • Memory integration functional
  • Understanding of memory flow
  • Production-ready architecture pattern

Quick Reference

MemMachine API Endpoints

  • Health: GET /api/v2/health
  • Add Memory: POST /api/v2/memories
  • Search Memories: POST /api/v2/memories/search
  • List Memories: POST /api/v2/memories/list

Key Environment Variables

MEMORY_SERVER_URL=https://YOUR_API_GATEWAY_URL
ORG_ID=workshop-org
PROJECT_ID=workshop-project
USER_ID=workshop-user
AWS_REGION=us-west-2
BEDROCK_MODEL_ID=openai.gpt-oss-20b-1:0

CloudFormation Template Parameters

StackName              - Name for the stack (default: memmachine-bedrock-ec2-stack)
InstanceType           - EC2 instance type (default: t3.xlarge)
KeyPairName            - EC2 key pair for SSH access
PostgresPassword       - PostgreSQL password (min 8 chars)
Neo4jPassword          - Neo4j password (min 8 chars)
AwsAccessKeyId         - AWS Access Key ID for Bedrock
AwsSecretAccessKey     - AWS Secret Access Key for Bedrock
AwsRegion              - AWS region for Bedrock (default: us-west-2)
BedrockEmbeddingModel  - Embedding model ID (default: amazon.titan-embed-text-v2:0)
BedrockLanguageModel   - Language model ID (default: openai.gpt-oss-20b-1:0)
AllowedCIDR            - CIDR for SSH/Neo4j access (default: 0.0.0.0/0)

Useful Commands

# Deploy stack
aws cloudformation create-stack \
    --stack-name memmachine-workshop \
    --template-body file://bedrock-cft.yml \
    --parameters ParameterKey=KeyPairName,ParameterValue=YOUR_KEY \
                 ParameterKey=PostgresPassword,ParameterValue=YOUR_PW \
                 ParameterKey=Neo4jPassword,ParameterValue=YOUR_PW \
                 ParameterKey=AwsAccessKeyId,ParameterValue=YOUR_KEY_ID \
                 ParameterKey=AwsSecretAccessKey,ParameterValue=YOUR_SECRET \
                 ParameterKey=AwsRegion,ParameterValue=us-west-2 \
    --capabilities CAPABILITY_NAMED_IAM \
    --region us-west-2

# Check status
aws cloudformation describe-stacks --stack-name YOUR_STACK

# Get outputs
aws cloudformation describe-stacks --stack-name YOUR_STACK --query 'Stacks[0].Outputs'

# Test API via API Gateway
curl https://YOUR_API_GATEWAY_URL/api/v2/health

Next Steps After Workshop

  • Review production considerations (security, scaling)
  • Explore advanced features (ProfileMemory, custom prompts)
  • Customize the Streamlit web interface
  • Integrate with your own applications
  • Join MemMachine community (Discord, GitHub)

Resources

Workshop Date: _____________ Your Name: _____________ Stack Name: _____________ API Gateway URL: _____________ Notes: _____________