llm-automation-docs-and-remediation-engine/deploy/helm/datacenter-docs/README.md

Datacenter Docs & Remediation Engine - Helm Chart

Helm chart for deploying the LLM Automation - Docs & Remediation Engine on Kubernetes.

Overview

This chart deploys a complete stack including:

• MongoDB: Document database for storing tickets, documentation, and metadata
• Redis: Cache and task queue backend
• API Service: FastAPI REST API with auto-remediation capabilities
• Chat Service: WebSocket server for real-time documentation queries (optional, not yet implemented)
• Worker Service: Celery workers for background tasks (optional, not yet implemented)
• Frontend: React-based web interface

Prerequisites

• Kubernetes 1.19+
• Helm 3.0+
• PersistentVolume provisioner support in the underlying infrastructure (for MongoDB persistence)
• Ingress controller (optional, for external access)

Installation

Quick Start

# Add the chart repository (if published)
helm repo add datacenter-docs https://your-repo-url
helm repo update

# Install with default values
helm install my-datacenter-docs datacenter-docs/datacenter-docs

# Or install from local directory
helm install my-datacenter-docs ./datacenter-docs

Production Installation

For production, create a custom values.yaml:

# Copy and edit the values file
cp values.yaml my-values.yaml

# Edit my-values.yaml with your configuration
# At minimum, change:
# - secrets.llmApiKey
# - secrets.apiSecretKey
# - ingress.hosts

# Install with custom values
helm install my-datacenter-docs ./datacenter-docs -f my-values.yaml

Install with Specific Configuration

helm install my-datacenter-docs ./datacenter-docs \
  --set secrets.llmApiKey="sk-your-openai-api-key" \
  --set secrets.apiSecretKey="your-strong-secret-key" \
  --set ingress.hosts[0].host="datacenter-docs.yourdomain.com" \
  --set mongodb.persistence.size="50Gi"

Configuration

Key Configuration Parameters

Global Settings

Parameter	Description	Default
global.imagePullPolicy	Image pull policy	IfNotPresent
global.storageClass	Storage class for PVCs	""

MongoDB

Parameter	Description	Default
mongodb.enabled	Enable MongoDB	true
mongodb.image.repository	MongoDB image	mongo
mongodb.image.tag	MongoDB version	7
mongodb.auth.rootUsername	Root username	admin
mongodb.auth.rootPassword	Root password	admin123
mongodb.persistence.enabled	Enable persistence	true
mongodb.persistence.size	Volume size	10Gi
mongodb.resources.requests.memory	Memory request	512Mi
mongodb.resources.limits.memory	Memory limit	2Gi

Redis

Parameter	Description	Default
redis.enabled	Enable Redis	true
redis.image.repository	Redis image	redis
redis.image.tag	Redis version	7-alpine
redis.resources.requests.memory	Memory request	128Mi
redis.resources.limits.memory	Memory limit	512Mi

API Service

Parameter	Description	Default
api.enabled	Enable API service	true
api.replicaCount	Number of replicas	2
api.image.repository	API image repository	datacenter-docs-api
api.image.tag	API image tag	latest
api.service.port	Service port	8000
api.autoscaling.enabled	Enable HPA	true
api.autoscaling.minReplicas	Min replicas	2
api.autoscaling.maxReplicas	Max replicas	10
api.resources.requests.memory	Memory request	512Mi
api.resources.limits.memory	Memory limit	2Gi

Worker Service

Parameter	Description	Default
worker.enabled	Enable worker service	false
worker.replicaCount	Number of replicas	3
worker.autoscaling.enabled	Enable HPA	true
worker.autoscaling.minReplicas	Min replicas	1
worker.autoscaling.maxReplicas	Max replicas	10

Chat Service

Parameter	Description	Default
chat.enabled	Enable chat service	false
chat.replicaCount	Number of replicas	1
chat.service.port	Service port	8001

Frontend

Parameter	Description	Default
frontend.enabled	Enable frontend	true
frontend.replicaCount	Number of replicas	2
frontend.service.port	Service port	80

Ingress

Parameter	Description	Default
ingress.enabled	Enable ingress	true
ingress.className	Ingress class	nginx
ingress.hosts[0].host	Hostname	datacenter-docs.example.com
ingress.tls[0].secretName	TLS secret name	datacenter-docs-tls

Application Configuration

Parameter	Description	Default
config.llm.baseUrl	LLM provider URL	https://api.openai.com/v1
config.llm.model	LLM model	gpt-4-turbo-preview
config.autoRemediation.enabled	Enable auto-remediation	true
config.autoRemediation.minReliabilityScore	Min reliability score	85.0
config.autoRemediation.dryRun	Dry run mode	false
config.logLevel	Log level	INFO

Secrets

Parameter	Description	Default
secrets.llmApiKey	LLM API key	sk-your-openai-api-key-here
secrets.apiSecretKey	API secret key	your-secret-key-here-change-in-production

IMPORTANT: Change these secrets in production!

Usage Examples

Enable All Services (including chat and worker)

helm install my-datacenter-docs ./datacenter-docs \
  --set chat.enabled=true \
  --set worker.enabled=true

Disable Auto-Remediation

helm install my-datacenter-docs ./datacenter-docs \
  --set config.autoRemediation.enabled=false

Use Different LLM Provider (e.g., Anthropic Claude)

helm install my-datacenter-docs ./datacenter-docs \
  --set config.llm.baseUrl="https://api.anthropic.com/v1" \
  --set config.llm.model="claude-3-opus-20240229" \
  --set secrets.llmApiKey="sk-ant-your-anthropic-key"

Use Local LLM (e.g., Ollama)

helm install my-datacenter-docs ./datacenter-docs \
  --set config.llm.baseUrl="http://ollama-service:11434/v1" \
  --set config.llm.model="llama2" \
  --set secrets.llmApiKey="not-needed"

Scale MongoDB Storage

helm install my-datacenter-docs ./datacenter-docs \
  --set mongodb.persistence.size="100Gi"

Disable Ingress (use port-forward instead)

helm install my-datacenter-docs ./datacenter-docs \
  --set ingress.enabled=false

Production Configuration with External MongoDB

# production-values.yaml
mongodb:
  enabled: false

config:
  mongodbUrl: "mongodb://user:pass@external-mongodb:27017/datacenter_docs?authSource=admin"

api:
  replicaCount: 5
  autoscaling:
    maxReplicas: 20

secrets:
  llmApiKey: "sk-your-production-api-key"
  apiSecretKey: "your-production-secret-key"

ingress:
  hosts:
    - host: "datacenter-docs.prod.yourdomain.com"
      paths:
        - path: /
          pathType: Prefix
          service: frontend
        - path: /api
          pathType: Prefix
          service: api

helm install prod-datacenter-docs ./datacenter-docs -f production-values.yaml

Upgrading

# Upgrade with new values
helm upgrade my-datacenter-docs ./datacenter-docs -f my-values.yaml

# Upgrade specific parameters
helm upgrade my-datacenter-docs ./datacenter-docs \
  --set api.image.tag="v1.2.0" \
  --reuse-values

Uninstallation

helm uninstall my-datacenter-docs

Note: This will delete all resources except PersistentVolumeClaims (PVCs) for MongoDB. To also delete PVCs:

kubectl delete pvc -l app.kubernetes.io/instance=my-datacenter-docs

Monitoring and Troubleshooting

Check Pod Status

kubectl get pods -l app.kubernetes.io/instance=my-datacenter-docs

View Logs

# API logs
kubectl logs -l app.kubernetes.io/component=api -f

# Worker logs
kubectl logs -l app.kubernetes.io/component=worker -f

# MongoDB logs
kubectl logs -l app.kubernetes.io/component=database -f

Access Services Locally

# API
kubectl port-forward svc/my-datacenter-docs-api 8000:8000

# Frontend
kubectl port-forward svc/my-datacenter-docs-frontend 8080:80

# MongoDB (for debugging)
kubectl port-forward svc/my-datacenter-docs-mongodb 27017:27017

Common Issues

Pods Stuck in Pending

Check if PVCs are bound:

kubectl get pvc

If storage class is missing, set it:

helm upgrade my-datacenter-docs ./datacenter-docs \
  --set mongodb.persistence.storageClass="standard" \
  --reuse-values

API Pods Crash Loop

Check logs:

kubectl logs -l app.kubernetes.io/component=api --tail=100

Common causes:

• MongoDB not ready (wait for init containers)
• Invalid LLM API key
• Missing environment variables

Cannot Access via Ingress

Check ingress status:

kubectl get ingress
kubectl describe ingress my-datacenter-docs

Ensure:

• Ingress controller is installed
• DNS points to ingress IP
• TLS certificate is valid (if using HTTPS)

Security Considerations

Production Checklist

• Change secrets.llmApiKey to a valid API key
• Change secrets.apiSecretKey to a strong random key
• Change MongoDB credentials (mongodb.auth.rootPassword)
• Enable TLS/SSL on ingress
• Review RBAC policies
• Use external secret management (e.g., HashiCorp Vault, AWS Secrets Manager)
• Enable network policies
• Set resource limits on all pods
• Enable pod security policies
• Review auto-remediation settings

Using External Secrets

Instead of storing secrets in values.yaml, use Kubernetes secrets:

# Create secret
kubectl create secret generic datacenter-docs-secrets \
  --from-literal=llm-api-key="sk-your-key" \
  --from-literal=api-secret-key="your-secret"

# Modify templates to use existing secret
# (requires chart customization)

Development

Validating the Chart

# Lint the chart
helm lint ./datacenter-docs

# Dry run
helm install my-test ./datacenter-docs --dry-run --debug

# Template rendering
helm template my-test ./datacenter-docs > rendered.yaml

Testing Locally

# Create kind cluster
kind create cluster

# Install chart
helm install test ./datacenter-docs \
  --set ingress.enabled=false \
  --set api.autoscaling.enabled=false \
  --set mongodb.persistence.enabled=false

# Test
kubectl port-forward svc/test-datacenter-docs-api 8000:8000
curl http://localhost:8000/health

Support

For issues and questions:

• Issues: https://git.commandware.com/it-ops/llm-automation-docs-and-remediation-engine/issues
• Documentation: https://git.commandware.com/it-ops/llm-automation-docs-and-remediation-engine

License

See the main repository for license information.