Truefoundry Docs

TrueFoundry provides a comprehensive platform for deploying Helm charts with support for multiple repository types, advanced configurations, and seamless integration with Kubernetes resources. This guide covers everything you need to know about deploying Helm charts through the TrueFoundry UI.

Overview

What You’ll Learn

This guide will walk you through:

Getting Started: How to access the Helm deployment interface and deploy your first chart
Repository Types: Understanding the three different ways to source Helm charts
Private Repositories: Setting up authentication for private chart repositories
Advanced Configuration: Customizing deployments with values, secrets, and Kustomize
Complete Example: A real-world Redis deployment with LoadBalancer exposure

Supported Repository Types

Helm charts in TrueFoundry can be deployed from three different source types:

HelmRepo: Traditional Helm repositories (HTTP/S servers hosting index.yaml files)
OCIRepo: OCI-compliant container registries (Docker Hub, GCR, ACR, etc.)
GitHelmRepo: Git repositories containing Helm charts (GitHub, GitLab, etc.)

Each repository type offers different advantages depending on your use case, from public chart repositories to private custom charts stored in Git.

Helm Chart Deployment Form

Deployment Process

Using the UI

Navigate to Deployments: Go to your workspace and click “New Deployment”
Select Helm: Choose “Helm” from the deployment options
Configure Repository: Select your repository type and fill in the details
Set Values: Configure chart values and advanced options
Deploy: Click “Submit” to deploy your chart

Repository Types

HelmRepo (Traditional Helm Repository)

OCIRepo (OCI Registry)

GitHelmRepo (Git Repository)

Private Repository Configuration

To use private repositories, you have to add and set up the repository along with credentials in the argocd repo secret. You can refer to the Private Repository Configuration guide to add and set up the repository along with credentials in the argocd repo secret.

Advanced Configuration

Cluster-scoped objects (like ClusterRole, ClusterRoleBinding, or non-namespace scoped resources) cannot be applied when deploying Helm charts unless you have cluster admin privileges. This restriction prevents workspace users from creating cluster-level resources that could impact other workloads.

Values Configuration

You can customize your Helm chart deployment using values files, these values will override the values in the chart.

Values Configuration

Two options available:

Values file as block file: Direct YAML input for chart values
Value files: Reference to values files in your repository (for GitHelmRepo)

Example values for Redis:

# Redis configuration
redis:
  auth:
    enabled: true
    password: "my-secure-password"

  master:
    persistence:
      enabled: true
      size: 8Gi

  replica:
    replicaCount: 2
    persistence:
      enabled: true
      size: 4Gi

# Service configuration
service:
  type: LoadBalancer
  ports:
    redis: 6379

Using Secrets in Helm Deployments

Kustomize Integration

TrueFoundry supports Kustomize for advanced customization of your Helm deployments:

Kustomize Configuration

Patch Configuration

Use patches to modify existing Kubernetes resources generated by your Helm chart:

patches:
  - target:
      group: apps
      version: v1
      kind: Deployment
      name: redis-master
    patch: |
      apiVersion: apps/v1
      kind: Deployment
      metadata:
        name: redis-master
      spec:
        template:
          metadata:
            annotations:
              prometheus.io/scrape: "true"
              prometheus.io/port: "9121"

Additional Manifests

Add custom Kubernetes resources alongside your Helm chart:

apiVersion: v1
kind: ConfigMap
metadata:
  name: redis-custom-config
  namespace: default
data:
  redis.conf: |
    maxmemory 2gb
    maxmemory-policy allkeys-lru
    save 900 1
    save 300 10
    save 60 10000

Exposing Endpoints with VirtualService

To expose your Helm chart deployment as an endpoint, you can add a VirtualService as an additional manifest:

apiVersion: networking.istio.io/v1beta1
kind: VirtualService
metadata:
  name: redis-virtualservice
  namespace: mitanshu-ws
spec:
  hosts:
    - redis-mitanshu-ws.ml.demo.truefoundry.cloud
  gateways:
    - istio-system/tfy-wildcard-alb
  tcp:
    - match:
        - port: 6379
      route:
        - destination:
            host: redis-redis-master.mitanshu-ws.svc.cluster.local
            port:
              number: 6379

Note: You’ll need to add this as an additional manifest in the Kustomize section to expose your Helm chart deployment as an endpoint.

Example: Exposing Redis Endpoint

Here’s a complete example of deploying Redis with a LoadBalancer service to expose the endpoint:

Step 1: Basic Configuration

Step 2: Create TrueFoundry Secret

Step 3: Values Configuration

# Values to expose Redis endpoint
values:
  auth:
    enabled: true
    password: tfy-secret://truefoundry:redis-secrets:password

  master:
    service:
      type: LoadBalancer
      ports:
        redis: 6379
    persistence:
      enabled: true
      size: 8Gi

  replica:
    replicaCount: 2
    service:
      type: ClusterIP
    persistence:
      enabled: true
      size: 4Gi

  metrics:
    enabled: true
    serviceMonitor:
      enabled: true

Step 4: Kustomize Patch and Endpoint Exposure (Optional)

Additional customization using Kustomize and exposing the endpoint by adding a VirtualService as an additional manifest:



# Additional manifest to expose Redis endpoint
apiVersion: networking.istio.io/v1beta1
kind: VirtualService
metadata:
  name: redis-virtualservice
  namespace: mitanshu-ws
spec:
  hosts:
    - redis-mitanshu-ws.ml.demo.truefoundry.cloud
  gateways:
    - istio-system/tfy-wildcard-alb
  tcp:
    - match:
        - port: 6379
      route:
        - destination:
            host: redis-redis-master.mitanshu-ws.svc.cluster.local
            port:
              number: 6379

Final Redis deployment showing the complete deployment with all configurations, Kustomize patches, and VirtualService applied

Final Redis Deployment Result

Getting Started

Train and Deploy Models

Service Deployment

Job Deployment

Workflow Deployment

Async Service Deployment

Volumes

ML Repository

Platform

Deploying On Your Own Cloud

Deploy Helm Charts

Overview

What You’ll Learn

Supported Repository Types

Deployment Process

Using the UI

Repository Types

Private Repository Configuration

Advanced Configuration

Create a Secret in TrueFoundry

Creating Kubernetes Secrets in Helm Deployments

1. Using Custom Kubernetes Objects

2. Using Kustomize

Patch Configuration

Additional Manifests

Exposing Endpoints with VirtualService

Example: Exposing Redis Endpoint

Getting Started

Train and Deploy Models

Service Deployment

Job Deployment

Workflow Deployment

Async Service Deployment

Volumes

ML Repository

Platform

Deploying On Your Own Cloud

​Overview

​What You’ll Learn

​Supported Repository Types

​Deployment Process

​Using the UI

​Repository Types

​Private Repository Configuration

​Advanced Configuration

​Example: Exposing Redis Endpoint

Overview

What You’ll Learn

Supported Repository Types

Deployment Process

Using the UI

Repository Types

Private Repository Configuration

Advanced Configuration

Example: Exposing Redis Endpoint