Masaic Dev Platform - Helm Chart

A Helm chart to deploy AgC(Agentic Compute)for an Agentic Orchestration AI platform.

Quick Installation

git clone https://github.com/your-org/platform.git
cd platform
helm upgrade --install -n masaic-cloud my-release ./masaic --create-namespace

Deployment Modes

1. Basic Setup (Default)

• In-memory DB, no vector store or external dependencies.

helm upgrade --install -n masaic-cloud my-release ./masaic --create-namespace

2. Production Setup

• Set the following variables in your values.yaml or with --set:
  • openResponses.qdrant.apikey (Qdrant API key)
  • openResponses.platform.model.apikey (Model API key)
  • openResponses.mongoDb.uri (MongoDB URI)
  • openResponses.vectorRepositoryType = mongodb
  • openResponses.vectorSearchProvider = qdrant
  • openResponses.storeType = mongodb
  • openResponses.otel.endpoint (OTEL collector endpoint)
  • openResponses.otel.headersMap (OTEL headers, e.g., API keys)

3. Built-in Secrets

• Use built-in secret management for API keys by setting:
  • openResponses.qdrant.apikey
  • openResponses.platform.model.apikey
  • openResponses.mongoDb.secrets.adminPassword

helm upgrade --install -n masaic-cloud my-release ./masaic --create-namespace \
  --set openResponses.qdrant.apikey="your_qdrant_api_key" \
  --set openResponses.platform.model.apikey="your_model_api_key" \
  --set openResponses.mongoDb.secrets.adminPassword="mongo_pwd"

Note: For third-party dependencies (MongoDB, SigNoz, Qdrant), see third-party installation guide.

Configuration

Environment Variables

Property	Description	Default	Options
openResponses.springProfile	Spring profile for the application	"platform"	Always set to "platform"
openResponses.vectorRepositoryType	Type of vector repository	"file"	"file", "mongodb"
openResponses.vectorSearchProvider	Vector search provider	"file"	"file", "qdrant"
openResponses.vectorSearchCollectionName	Name of the vector search collection	"openresponses"	Custom string
openResponses.storeType	Storage type for model messages	"in-memory"	"in-memory", "mongodb"
openResponses.mongoDbDatabase	MongoDB database name	"openresponses"	Custom string
openResponses.qdrant.host	Qdrant server host	""	Qdrant server URL
openResponses.qdrant.port	Qdrant server port	6334	Port number
openResponses.qdrant.apikey	Qdrant API key	""	API key string
openResponses.qdrant.useTLS	Enable TLS for Qdrant	true	true, false
openResponses.mongoDb.uri	MongoDB connection URI	""	MongoDB connection string
openResponses.platform.model.apikey	API key for the model provider	""	API key string
openResponses.platform.model.name	Model name to use	"gpt-4.1-mini"	Model identifier
openResponses.otel.sdkDisabled	Disable OTEL SDK	false	true, false
openResponses.otel.endpoint	OTEL collector endpoint	""	Collector URL
openResponses.otel.headersMap	OTEL headers (e.g., API keys)	{}	Key-value pairs
platformui.openResponses.url	URL for OpenResponses service	"http://localhost:6644"	Service URL

Service Ports

• OpenResponses API: Port 6644
• Platform UI: Port 6645

Health Endpoints

• OpenResponses: /actuator/health
• Platform UI: /health

Resource Limits Example

openResponses:
  resources:
    requests:
      cpu: 2
      memory: 1Gi
    limits:
      cpu: 2
      memory: 1Gi
platformui:
  resources:
    requests:
      cpu: 1
      memory: 512Mi
    limits:
      cpu: 1
      memory: 512Mi

Ingress Example

openResponses:
  ingress:
    enabled: true
    hosts:
      - host: api.yourdomain.com
        paths:
          - path: /
            pathType: ImplementationSpecific
platformui:
  ingress:
    enabled: true
    hosts:
      - host: ui.yourdomain.com
        paths:
          - path: /
            pathType: ImplementationSpecific

Troubleshooting

• Check pod status: kubectl get pods -n masaic-cloud
• View logs: kubectl logs -n masaic-cloud deployment/masaic-platform
• Check services: kubectl get svc -n masaic-cloud

Cleanup

To remove all resources created by this chart, run:

./scripts/cleanup.sh