# Deployment method configuration
deployment:
method: standard # Deployment method: "standard" (K8s resources) or "toolhive" (MCPServer CRD)
# Application image configuration
image:
repository: ghcr.io/vfarcic/dot-ai # Container image repository
tag: "0.125.0" # Container image tag - set by CI pipeline during release
# Resource configuration
resources:
requests:
memory: "512Mi" # Minimum memory required
cpu: "200m" # Minimum CPU required
limits:
memory: "2Gi" # Maximum memory allowed
cpu: "1000m" # Maximum CPU allowed
# Secrets configuration
secrets:
name: dot-ai-secrets # Name of the Kubernetes Secret resource
anthropic:
keyName: anthropic-api-key # Key name within the secret
apiKey: "" # API key value (only needed if chart should create the secret)
openai:
keyName: openai-api-key # Key name within the secret
apiKey: "" # API key value (only needed if chart should create the secret)
google:
keyName: google-api-key # Key name within the secret
apiKey: "" # API key value (only needed if chart should create the secret)
xai:
keyName: xai-api-key # Key name within the secret
apiKey: "" # API key value (only needed if chart should create the secret)
mistral:
keyName: mistral-api-key # Key name within the secret
apiKey: "" # API key value (only needed if chart should create the secret)
deepseek:
keyName: deepseek-api-key # Key name within the secret
apiKey: "" # API key value (only needed if chart should create the secret)
# ServiceAccount configuration
serviceAccount:
create: true # Create a ServiceAccount
name: "" # ServiceAccount name override (generated if empty)
# Ingress configuration
ingress:
enabled: false # Create Ingress resource
className: nginx # Ingress class name
host: dot-ai.127.0.0.1.nip.io # Ingress hostname
# Annotations required for HTTP transport with SSE (Server-Sent Events)
# If using different className, update annotations for your ingress controller:
# - Traefik: traefik.ingress.kubernetes.io/service.sticky.cookie.httponly: "true"
# - HAProxy: haproxy.org/timeout-http-request: "3600s"
# - AWS ALB: alb.ingress.kubernetes.io/target-group-attributes: idle_timeout.timeout_seconds=3600
annotations:
nginx.ingress.kubernetes.io/proxy-read-timeout: "3600" # Allow long-running SSE connections
nginx.ingress.kubernetes.io/proxy-send-timeout: "3600" # Allow long-running SSE connections
nginx.ingress.kubernetes.io/proxy-buffering: "off" # Disable buffering for real-time streaming
nginx.ingress.kubernetes.io/proxy-request-buffering: "off" # Disable request buffering for real-time streaming
tls:
enabled: false # Enable TLS/HTTPS
secretName: "" # TLS secret name (generated if empty when enabled)
# AI Provider configuration
ai:
provider: anthropic # AI provider type (anthropic, openai, google, xai, mistral, deepseek)
model: "" # Optional: model override (e.g., "llama3.1:70b", "gpt-4o")
# Custom endpoint configuration for self-hosted or alternative SaaS providers (PRD #194)
customEndpoint:
enabled: false # Enable custom endpoint
baseURL: "" # Custom LLM endpoint URL (e.g., "http://ollama-service:11434/v1")
embeddingsBaseURL: "" # Optional: custom embeddings endpoint URL (if different from LLM endpoint)
# Examples (commented out):
# Example 1: Ollama (self-hosted)
# ai:
# provider: openai
# model: "llama3.1:70b"
# customEndpoint:
# enabled: true
# baseURL: "http://ollama-service:11434/v1"
#
# Example 2: Azure OpenAI (SaaS)
# ai:
# provider: openai
# model: "gpt-4o"
# customEndpoint:
# enabled: true
# baseURL: "https://YOUR_RESOURCE.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT"
#
# Example 3: vLLM (self-hosted)
# ai:
# provider: openai
# model: "meta-llama/Llama-3.1-70B-Instruct"
# customEndpoint:
# enabled: true
# baseURL: "http://vllm-service:8000/v1"
#
# Note: Custom endpoints must support OpenAI-compatible API and models must
# support 8K+ output tokens for reliable YAML generation.
# Qdrant Vector Database
qdrant:
enabled: true # Deploy Qdrant as dependency (false = use external)
image:
repository: qdrant/qdrant # Qdrant image repository
tag: v1.15.5 # Qdrant image tag
external:
url: "" # External Qdrant URL (required when enabled=false)
apiKey: "" # External Qdrant API key (optional)
