DevOps AI Toolkit

# 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: "0.156.0" # 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 auth: keyName: auth-token # Key name within the secret for Bearer token auth token: "" # Auth token value (only needed if chart should create the secret) 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) moonshot: keyName: moonshot-api-key # Key name within the secret (PRD #237: Kimi K2) apiKey: "" # API key value (only needed if chart should create the secret) customLlm: keyName: custom-llm-api-key # Key name within the secret for custom LLM endpoint apiKey: "" # API key value (only needed if chart should create the secret) customEmbeddings: keyName: custom-embeddings-api-key # Key name within the secret for custom embeddings endpoint 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) clusterIssuer: "" # cert-manager ClusterIssuer name (e.g., "letsencrypt") # AI Provider configuration ai: provider: anthropic # AI provider type (anthropic, anthropic_opus, anthropic_haiku, openai, google, kimi, kimi_thinking, xai, amazon_bedrock) 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 - MUST include /v1 suffix for OpenAI-compatible APIs (e.g., "http://ollama-service:11434/v1") embeddingsBaseURL: "" # Optional: custom embeddings endpoint URL (if different from LLM endpoint, also requires /v1 suffix) embeddingsModel: "" # Optional: custom embeddings model name (e.g., "nomic-embed-text" for Ollama, defaults to "text-embedding-3-small" for OpenAI) embeddingsDimensions: "" # Optional: custom embeddings dimensions (e.g., "768" for nomic-embed-text, defaults to "1536" for OpenAI) # Examples (commented out): # Example 1: Ollama (self-hosted) - IMPORTANT: Include /v1 suffix # ai: # provider: openai # model: "llama3.1:70b" # customEndpoint: # enabled: true # baseURL: "http://ollama-service:11434/v1" # /v1 suffix is REQUIRED # # 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) - IMPORTANT: Include /v1 suffix # ai: # provider: openai # model: "meta-llama/Llama-3.1-70B-Instruct" # customEndpoint: # enabled: true # baseURL: "http://vllm-service:8000/v1" # /v1 suffix is REQUIRED # # Note: OpenAI-compatible endpoints (Ollama, vLLM, LocalAI) REQUIRE the /v1 suffix. # Without it, API calls will fail with 404 Not Found errors. # Custom endpoints must support OpenAI-compatible API and models must # support 8K+ output tokens for reliable YAML generation. # Additional environment variables (optional) # Use this to add any custom environment variables to the MCP server # Example use cases: tracing configuration, custom integrations, feature flags extraEnv: [] # - name: OTEL_TRACING_ENABLED # value: "true" # - name: OTEL_EXPORTER_OTLP_ENDPOINT # value: "http://jaeger-collector:4318/v1/traces" # - name: OTEL_SERVICE_NAME # value: "dot-ai-mcp-production" # dot-ai Controller Solution CR # Note: Controller must be installed separately before enabling this # Install: helm install dot-ai-controller oci://ghcr.io/vfarcic/dot-ai-controller/charts/dot-ai-controller:0.14.0 -n dot-ai controller: enabled: false # Create Solution CR to track this deployment (requires controller installed separately) # 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)