Skip to content

Latest commit

 

History

History
281 lines (195 loc) · 7.87 KB

CLOUD_FOUNDRY.md

File metadata and controls

281 lines (195 loc) · 7.87 KB

Cloud Foundry Deployment Guide

This guide provides detailed instructions for deploying the Movie Chatbot application to Cloud Foundry, with a focus on configuration options and service bindings.

Table of Contents

Prerequisites

Before deploying the application, ensure you have:

  • Cloud Foundry CLI installed (cf command available)
  • Access to a Tanzu Platform for Cloud Foundry environment
  • Proper permissions to create and bind services
  • The application code from the repository

Configuration Options

The application supports multiple configuration sources, with the following priority order:

  1. Service Bindings (highest priority)

    • GenAI tile service bindings for LLM configuration
    • User-defined services for other required/optional configuration

  2. Environment Variables (second priority)

    • Set via cf set-env or in manifest.yml
    • Used when service bindings are not available

  3. config.json (third priority)

    • Used as a fallback when neither service bindings nor environment variables are available
    • Particularly useful for local development or simplified deployment

Deployment Scenarios

Scenario 1: With GenAI Tile Service Binding

This is the recommended approach for production deployments when the GenAI tile is available.

  1. Deploy the application:

    # Clone the repository if you haven't already
git clone https://github.com/cf-toolsuite/tanzu-genai-showcase
cd tanzu-genai-showcase/py-django-crewai

# Deploy to Cloud Foundry
./deploy-on-tp4cf.sh

  2. Create and bind a GenAI service instance:

    # Discover available GenAI tile service offering plans
cf marketplace -e genai

# Create a GenAI service instance
cf create-service genai PLAN_NAME movie-chatbot-llm

# Bind the application to the service
cf bind-service movie-chatbot movie-chatbot-llm

  3. Set required non-LLM configuration:

    cf set-env movie-chatbot DJANGO_SECRET_KEY "your_secret_key"
cf set-env movie-chatbot TMDB_API_KEY "your_tmdb_api_key"

  4. Start the application:

    cf start movie-chatbot

Scenario 2: With User-Defined Services

This approach is useful when the GenAI tile is not available, but you still want to use service bindings for configuration.

  1. Deploy the application:

    ./deploy-on-tp4cf.sh

  2. Create a user-defined service for LLM configuration:

    cf create-user-provided-service movie-chatbot-llm -p '{
  "api_key": "your_llm_api_key",
  "api_base": "your_llm_base_url",
  "model_name": "your_model_name",
  "model_provider": "openai"
}'

  3. Create a user-defined service for other required configuration:

    cf create-user-provided-service movie-chatbot-config -p '{
  "DJANGO_SECRET_KEY": "your_secret_key",
  "TMDB_API_KEY": "your_tmdb_api_key",
  "SERPAPI_API_KEY": "your_serpapi_key"
}'

  4. Bind the services to the application:

    cf bind-service movie-chatbot movie-chatbot-llm
cf bind-service movie-chatbot movie-chatbot-config

  5. Start the application:

    cf start movie-chatbot

Scenario 3: With Environment Variables

This approach uses environment variables for all configuration.

  1. Deploy the application:

    ./deploy-on-tp4cf.sh

  2. Set LLM configuration:

    cf set-env movie-chatbot OPENAI_API_KEY "your_llm_api_key"
cf set-env movie-chatbot LLM_BASE_URL "your_llm_base_url"
cf set-env movie-chatbot LLM_MODEL "your_model_name"

  3. Set other required configuration:

    cf set-env movie-chatbot DJANGO_SECRET_KEY "your_secret_key"
cf set-env movie-chatbot TMDB_API_KEY "your_tmdb_api_key"

  4. Set optional configuration (if needed):

    cf set-env movie-chatbot SERPAPI_API_KEY "your_serpapi_key"
cf set-env movie-chatbot MOVIE_RESULTS_LIMIT "5"
cf set-env movie-chatbot MAX_RECOMMENDATIONS "3"

  5. Start the application:

    cf start movie-chatbot

Scenario 4: With config.json

This approach uses a config.json file for configuration, which is included in the deployment.

  1. Create a config.json file:

    {
  "DJANGO_SECRET_KEY": "your_secret_key",
  "TMDB_API_KEY": "your_tmdb_api_key",
  "OPENAI_API_KEY": "your_llm_api_key",
  "LLM_BASE_URL": "your_llm_base_url",
  "LLM_MODEL": "your_model_name",
  "SERPAPI_API_KEY": "your_serpapi_key",
  "MOVIE_RESULTS_LIMIT": "5",
  "MAX_RECOMMENDATIONS": "3"
}

  2. Deploy the application with the config.json file:

    ./deploy-on-tp4cf.sh

  3. Start the application:

    cf start movie-chatbot

Required vs. Optional Configuration

Required Configuration

These settings are essential for the application to function:

  • DJANGO_SECRET_KEY: Required for Django security features
  • TMDB_API_KEY: Required for movie data retrieval

LLM Configuration (Required for AI functionality)

These settings are specifically for the LLM service:

  • LLM API Key: Required for LLM access (via OPENAI_API_KEY or LLM_API_KEY)
  • LLM Base URL: Optional, defaults to OpenAI's API endpoint
  • LLM Model: Optional, defaults to 'gpt-4o-mini'

Optional Configuration

These settings have defaults but can be overridden:

  • SERPAPI_API_KEY: Optional, for theater showtimes (enhances functionality)
  • MOVIE_RESULTS_LIMIT: Optional, defaults to 5
  • MAX_RECOMMENDATIONS: Optional, defaults to 3
  • THEATER_SEARCH_RADIUS_MILES: Optional, defaults to 15
  • DEFAULT_SEARCH_START_YEAR: Optional, defaults to 1900
  • Other API request configuration parameters (timeouts, retries, etc.)

Service Binding Details

GenAI Tile Service Binding

When using the GenAI tile, the application automatically extracts the following from the service binding:

  • api_key: The API key for the LLM service
  • api_base: The base URL for the LLM service
  • model_name: The name of the model to use
  • model_provider: The provider of the model (e.g., 'openai')

The application looks for services with the 'genai' tag or label, or specifically named 'movie-chatbot-llm'.

User-Defined Services

When using user-defined services, you can create two types of services:

  1. LLM Service (named 'movie-chatbot-llm'):

    • Contains LLM-specific configuration (api_key, api_base, model_name, model_provider)

  2. Config Service (named 'movie-chatbot-config'):

    • Contains other required and optional configuration (DJANGO_SECRET_KEY, TMDB_API_KEY, etc.)

Troubleshooting

Common Issues

  1. Missing Required Configuration:

    • Check that DJANGO_SECRET_KEY and TMDB_API_KEY are set
    • Verify that LLM API key is available (via service binding or environment variable)

  2. Service Binding Issues:

    • Verify that services are correctly bound to the application
    • Check service binding credentials with cf env movie-chatbot

  3. Configuration Priority:

    • Remember that service bindings take precedence over environment variables
    • Environment variables take precedence over config.json

Debugging

To enable debug logging:

cf set-env movie-chatbot DEBUG True
cf set-env movie-chatbot LOG_LEVEL DEBUG
cf restage movie-chatbot

View logs:

cf logs movie-chatbot --recent