Skip to content

chat-history-configuration.md

Latest commit

 

History

History
348 lines (256 loc) · 8.11 KB

chat-history-configuration.md

File metadata and controls

348 lines (256 loc) · 8.11 KB

Chat History Configuration Guide

This guide explains how to configure and use the different chat history providers available in the Partnership Agent.

Overview

The Partnership Agent supports three chat history storage options:

  • InMemory - Default, no persistence, fastest startup
  • SQLite - Local persistence with Docker container
  • Azure SQL - Cloud persistence for production scenarios

Quick Setup

Using Setup Scripts (Recommended)

# InMemory (default)
./setup/setup.sh

# SQLite with persistence
./setup/setup.sh sqlite

# Azure SQL Database
./setup/setup.sh azuresql

Manual Configuration

cd setup
dotnet run --project CrossPlatformSetup.csproj [inmemory|sqlite|azuresql]

Provider Details

🧠 InMemory Provider

Best for: Development, testing, quick demos

Configuration: No additional configuration required

Pros:

  • Zero setup time
  • Fastest performance
  • No external dependencies

Cons:

  • No persistence - data lost on restart
  • Not suitable for production

Usage:

./setup/setup.sh inmemory
# or simply
./setup/setup.sh

🗄️ SQLite Provider

Best for: Local development with persistence

Configuration: Automatic Docker container setup

Pros:

  • Full persistence
  • Local file-based storage
  • No cloud dependencies
  • Automatic container management

Cons:

  • Requires Docker
  • Single-user scenarios only

Container Details:

  • Container name: partnership-agent-sqlite
  • Database file: /data/partnership-agent.db
  • Volume: partnership-agent-sqlite-data

Usage:

./setup/setup.sh sqlite

Manual Container Management:

# Check container status
docker ps -a --filter name=partnership-agent-sqlite

# View container logs
docker logs partnership-agent-sqlite

# Stop container
docker stop partnership-agent-sqlite

# Remove container and volume
docker rm partnership-agent-sqlite
docker volume rm partnership-agent-sqlite-data

☁️ Azure SQL Provider

Best for: Production deployments, multi-user scenarios

Configuration: Requires Azure SQL Database setup

Pros:

  • Enterprise-grade reliability
  • Automatic backups
  • Multi-user support
  • Scalable performance

Cons:

  • Requires Azure subscription
  • Network dependency
  • More complex setup

Prerequisites:

  1. Azure SQL Database provisioned
  2. Connection string configured
  3. Database tables created

Setup Steps:

  1. Configure connection string:

    cd src/PartnershipAgent.WebApi
dotnet user-secrets set "AzureSQL:ConnectionString" "Server=tcp:your-server.database.windows.net,1433;Initial Catalog=your-database;Persist Security Info=False;User ID=your-username;Password=your-password;MultipleActiveResultSets=False;Encrypt=True;TrustServerCertificate=False;Connection Timeout=30;"

  2. Run setup:

    ./setup/setup.sh azuresql

  3. Verify database table creation: The setup automatically creates the ChatMessages table with the required schema.

Configuration Methods

1. Command-Line Arguments (Recommended)

The setup scripts accept provider arguments:

./setup/setup.sh [inmemory|sqlite|azuresql]

2. User Secrets

Configure the provider and connection strings using .NET user secrets:

cd src/PartnershipAgent.WebApi

# Set provider
dotnet user-secrets set "ChatHistory:Provider" "sqlite"

# Set SQLite connection (if using SQLite)
dotnet user-secrets set "SQLite:ConnectionString" "Data Source=/data/partnership-agent.db;Cache=Shared"

# Set Azure SQL connection (if using Azure SQL)
dotnet user-secrets set "AzureSQL:ConnectionString" "your-connection-string"

3. Configuration Files

Modify appsettings.json or appsettings.Development.json:

{
  "ChatHistory": {
    "Provider": "sqlite"
  },
  "SQLite": {
    "ConnectionString": "Data Source=/data/partnership-agent.db;Cache=Shared"
  },
  "AzureSQL": {
    "ConnectionString": "your-connection-string"
  }
}

Database Schema

All providers use the same chat message schema:

CREATE TABLE ChatMessages (
    Id UNIQUEIDENTIFIER PRIMARY KEY,
    ThreadId NVARCHAR(100) NOT NULL,
    Role NVARCHAR(50) NOT NULL,
    Content NVARCHAR(MAX) NOT NULL,
    CreatedAt DATETIME2 NOT NULL,
    UserId NVARCHAR(100),
    TenantId NVARCHAR(100)
);

CREATE INDEX IX_ChatMessages_ThreadId ON ChatMessages(ThreadId);
CREATE INDEX IX_ChatMessages_UserId ON ChatMessages(UserId);
CREATE INDEX IX_ChatMessages_TenantId ON ChatMessages(TenantId);
CREATE INDEX IX_ChatMessages_CreatedAt ON ChatMessages(CreatedAt);

Switching Providers

During Development

Stop the application and run setup with the new provider:

# Stop current application (Ctrl+C)

# Switch to SQLite
./setup/setup.sh sqlite

# Or switch to Azure SQL
./setup/setup.sh azuresql

Runtime Configuration

Update the configuration and restart the application:

# Update user secrets
dotnet user-secrets set "ChatHistory:Provider" "azuresql"

# Restart application
cd src/PartnershipAgent.WebApi
dotnet run

Troubleshooting

SQLite Issues

Container not starting:

# Check Docker status
docker ps -a --filter name=partnership-agent-sqlite

# View logs
docker logs partnership-agent-sqlite

# Recreate container
docker rm partnership-agent-sqlite
docker volume rm partnership-agent-sqlite-data
./setup/setup.sh sqlite

Database file permissions:

  • Ensure Docker has access to create volumes
  • Check Docker Desktop settings for file sharing

Azure SQL Issues

Connection failures:

  • Verify connection string format
  • Check firewall rules in Azure portal
  • Ensure database exists and is accessible

Authentication errors:

  • Verify username/password in connection string
  • Check Azure AD authentication if applicable
  • Ensure user has appropriate database permissions

Table creation errors:

  • Verify user has CREATE TABLE permissions
  • Check if tables already exist
  • Run table creation script manually if needed

General Issues

Provider not recognized:

  • Check spelling: inmemory, sqlite, azuresql
  • Verify configuration is properly set
  • Check application logs for configuration errors

Performance issues:

  • InMemory: Fastest, no persistence
  • SQLite: Good performance, local storage
  • Azure SQL: Network dependent, check latency

Testing Chat History

Verify Persistence

  1. Start application with chosen provider
  2. Send a chat message via API: 
    curl -X POST "http://localhost:5001/api/chat" \
  -H "Content-Type: application/json" \
  -d '{
    "threadId": "test-persistence",
    "prompt": "Hello, test message"
  }'
  3. Restart application
  4. Send another message with same threadId
  5. Verify conversation history is maintained

Load Testing

For production scenarios with Azure SQL:

# Multiple concurrent requests
for i in {1..10}; do
  curl -X POST "http://localhost:5001/api/chat" \
    -H "Content-Type: application/json" \
    -d "{
      \"threadId\": \"load-test-$i\",
      \"prompt\": \"Load test message $i\"
    }" &
done
wait

Best Practices

Development

  • Use InMemory for quick testing and development
  • Use SQLite when you need to test persistence locally
  • Keep SQLite containers running during development sessions

Production

  • Use Azure SQL for production deployments
  • Configure proper connection pooling
  • Set up monitoring and alerts
  • Regular database backups
  • Consider read replicas for high-load scenarios

Security

  • Never commit connection strings to source control
  • Use Azure Key Vault for production secrets
  • Implement proper authentication and authorization
  • Regular security updates for database engines

Performance Considerations

Provider Startup Time Request Latency Persistence Concurrency
InMemory Fastest Lowest None Single Process
SQLite Fast Low Full Single User
Azure SQL Moderate Variable Full Multi-User

Choose the provider that best matches your performance and persistence requirements.