Butler AI Chat App

Butler is an AI chat app, which covers all fronts of development, from frontend to backend, and from design to deployment.

Installation

Backend Orchestration

The backend is organized into microservices. Each microservice is a separate project and has a Dockerfile to build a Docker image. Kubernetes can be used to deploy the Docker images.

Secrets

You have to create a butler-secret.yaml kind: Secret file with the following keys and values:

data:
    JWT_SECRET: cGFzc3dvcmQ= # "password" in base64
    POSTGRES_PASSWORD: cGFzc3dvcmQ= # "password" in base64
    REDIS_PASSWORD: cGFzc3dvcmQ= # "password" in base64
    PGADMIN_DEFAULT_PASSWORD: cGFzc3dvcmQ= # "password" in base64

Deployment to Kubernetes

Run the following commands to build the Docker images and deploy them to Kubernetes:

docker context use default
# Start minikube (should set kubectl to use minikube's context)
minikube start --listen-address=0.0.0.0
minikube addons enable ingress

minikube docker-env
# Set the Docker environment variables to the minikube Docker daemon
eval $(minikube docker-env)
# or `& minikube -p minikube docker-env --shell powershell | Invoke-Expression` for Windows PowerShell

# Build the Docker images (in the correct minikube docker-env)
docker-compose build

# Load images

# Deploy the config files to Kubernetes
kubectl apply -f butler-configmap.yaml
kubectl apply -f butler-secret.yaml # You have to create this file
kubectl apply -f api_gateway/deployment.yaml
kubectl apply -f api_gateway/service.yaml
# kubectl apply -f services/localai/deployment.yaml # Uncomment if you want self-hosted AI
kubectl apply -f services/postgresql/postgresql-statefulset.yaml
kubectl apply -f services/postgresql/pgadmin-deployment.yaml
kubectl apply -f services/redis/redis-configmap.yaml
kubectl apply -f services/redis/redis-statefulset.yaml
kubectl apply -f services/redis/sentinel-statefulset.yaml
kubectl apply -f services/chat_service/deployment.yaml
kubectl apply -f services/chat_service/service.yaml
kubectl apply -f services/identity_service/deployment.yaml
kubectl apply -f services/identity_service/service.yaml
kubectl apply -f services/ai_service/deployment.yaml
kubectl apply -f services/ai_service/service.yaml

# Separate namespace for monitoring
kubectl create namespace monitoring
kubectl apply -f services/prometheus/clusterRole.yaml
kubectl apply -f services/prometheus/config-map.yaml
kubectl apply -f services/prometheus/prometheus-deployment.yaml
kubectl apply -f services/prometheus/prometheus-service.yaml
kubectl apply -f services/grafana/grafana-datasource-config.yaml
kubectl apply -f services/grafana/deployment.yaml
kubectl apply -f services/grafana/service.yaml

# For Jaeger and OpenTelemetry
# Install cert manager needed for Jaeger
kubectl create namespace observability
kubectl apply -f https://github.com/cert-manager/cert-manager/releases/download/v1.14.5/cert-manager.yaml
kubectl apply -n observability -f https://github.com/jaegertracing/jaeger-operator/releases/download/v1.57.0/jaeger-operator.yaml
# 5 second delay in PowerShell
Start-Sleep -Seconds 5.0
kubectl apply -n observability -f services/jaeger/simplest.yaml

To expose the API Gateway's and pgAdmin's port, use command

minikube service butler-api-gateway-service
minikube service pgadmin
minikube service prometheus-service --namespace monitoring
minikube service grafana --namespace monitoring
minikube service jaeger-spm-query --namespace observability

Exposing services

To expose the API Gateway's port, use the following command:

kubectl port-forward --address 0.0.0.0 service/butler-api-gateway-service 8080:8080 # Exposing the API Gateway's port

Now, port 8080 is exposed to the host machine. You can access the API Gateway at http://<your_host_ip_address>:8080 on LAN. With port-forwarding enabled on your router, you can access the API Gateway from anywhere.

Monitoring

To monitor PostgreSQL, check the exposed url got from running minikube service pgadmin and login with the credentials, found in butler-configmap.yaml and butler-secret.yaml (created by you). Then you can connect to your database on port 5432 with the credentials found in butler-secret.yaml. Keep in mind, that you may have to modify the PostgreSQL superuser in a pod's shell with the following command:

psql -U postgres
\password postgres
# Enter the new password

Frontend

To build and run the Compose Multiplatform app, you should have a local.properties file in the client directory with the following content:

RELEASE_KEYSTORE_PASSWORD=your_release_keystore_password
RELEASE_KEY_PASSWORD=your_release_key_password
RELEASE_KEY_ALIAS=your_release_key_alias
RELEASE_KEY_PATH=path/to/release.keystore

DEBUG_KEYSTORE_PASSWORD=your_debug_keystore_password
DEBUG_KEY_PASSWORD=your_debug_key_password
DEBUG_KEY_ALIAS=your_debug_key_alias
DEBUG_KEY_PATH=path/to/debug.keystore

API_GATEWAY_URL=http://127.0.0.1:12345

Usage

Contributing

See more information in CONTRIBUTING.md.

To reset deployments and images, the following commands should help:

# Delete the deployments
kubectl delete -n default deployment butler-api-gateway
kubectl delete -n default deployment butler-chat-service
kubectl delete -n default deployment butler-identity-service
kubectl delete -n default deployment butler-ai-service

# Delete the images
minikube image rm illyan1337/butler-api-gateway:latest
minikube image rm illyan1337/butler-chat-service:latest
minikube image rm illyan1337/butler-identity-service:latest
minikube image rm illyan1337/butler-ai-service:latest

# Build the images again
docker-compose build

# Redeploy the microservices
kubectl apply -f api_gateway/deployment.yaml
kubectl apply -f services/chat_service/deployment.yaml
kubectl apply -f services/identity_service/deployment.yaml
kubectl apply -f services/ai_service/deployment.yaml

License

GPLv3