Shop Platform Helm Chart

Reusable Helm chart for deploying multiple application components with shared structure and environment-based configuration.

Supports:

• Multiple components
• Multiple instances per component (production fan-out)
• Shared global configuration
• Per-component shared config
• Per-instance overrides
• Resource inheritance and overrides

---

Architecture

Global
  ↓
Component
  ↓
Instance

Configuration order (highest override wins):

global → component → instance

---

Directory Structure

shop-platform/

├── Chart.yaml
├── templates/
│   ├── deployment.yaml
│   └── configmap.yaml
│
└── values/
    ├── dev.yaml
    ├── preprod.yaml
    └── prod.yaml

---

Installation

Dev

helm upgrade --install shop-dev . \
  -f values/dev.yaml

Preprod

helm upgrade --install shop-preprod . \
  -f values/preprod.yaml

Prod

helm upgrade --install shop-prod . \
  -f values/prod.yaml

---

Values Structure

global:

components:
  - componentName:
    image:
    replicas:
    resources:
    commonConfig:
    instances:

---

Global Configuration

Applies to all deployments generated by the chart.

global:
  annotations:
    prometheus.io/scrape: "true"

  labels:
    team: backend

  imagePullSecrets:
    - name: prod-regcred

  resources:
    requests:
      cpu: 100m
      memory: 128Mi
    limits:
      cpu: 500m
      memory: 512Mi

  nodeSelector:
    workload: backend

Supported global fields

Field	Description
annotations	Applied to all deployments/pods
labels	Shared labels
imagePullSecrets	Registry credentials
resources	Default resource limits/requests
nodeSelector	Scheduling constraints
tolerations	Scheduling constraints
affinity	Scheduling constraints

---

Adding a Component

Example: inventory

components:
  - componentName: inventory

    image: ghcr.io/acme/inventory-service:1.0.0

    replicas: 2

    commonConfig:
      DB_PORT: "5432"
      LOG_LEVEL: info

    instances:
      - name: default
        config:
          DB_HOST: inventory-db

---

Multiple Instances per Component

Example: orders

components:
  - componentName: orders

    image: ghcr.io/acme/orders-service:1.3.0

    replicas: 5

    commonConfig:
      DB_PORT: "5432"
      LOG_LEVEL: warn

    instances:
      - name: eu
        config:
          DB_HOST: postgres-eu
          REGION: eu

      - name: us
        config:
          DB_HOST: postgres-us
          REGION: us

      - name: asia
        config:
          DB_HOST: postgres-asia
          REGION: asia

Generated:

orders-eu
orders-us
orders-asia

---

ConfigMap Behavior

ConfigMap = merge of:

component.commonConfig
+ instance.config

Example:

commonConfig:
  DB_PORT: "5432"
  LOG_LEVEL: warn

instance:
  config:
    DB_HOST: postgres-eu

Result:

data:
  DB_PORT: "5432"
  LOG_LEVEL: "warn"
  DB_HOST: "postgres-eu"

---

Resource Handling

Resources are resolved in this order:

global.resources → component.resources → (override)

Global default

global:
  resources:
    requests:
      cpu: 100m
      memory: 128Mi
    limits:
      cpu: 500m
      memory: 512Mi

Component override

components:
  - componentName: orders

    resources:
      requests:
        cpu: 1
        memory: 1Gi
      limits:
        cpu: 4
        memory: 4Gi

---

Environment Strategy

dev

• Single instance per component
• Minimal resources
• Debug logging

preprod

• Production-like setup
• Slightly reduced replicas
• Stable images (RC tags)

prod

• Full scaling
• Multi-instance components (e.g. orders-eu/us/asia)
• Region-based configuration

---

Example Production Layout

gateway
orders-eu
orders-us
orders-asia
payments
notifications

---

Adding a New Environment

1. Create a new file:

values/staging.yaml

2. Copy structure from preprod or dev

3. Adjust:

• images
• replicas
• instance definitions

Run:

helm upgrade --install shop-staging . \
  -f values/staging.yaml

---

Common Commands

Lint:

helm lint .

Render:

helm template shop . -f values/prod.yaml

Deploy:

helm upgrade --install shop . -f values/prod.yaml

Rollback:

helm rollback shop 1

---

Best Practices

• Keep environment differences ONLY in values/*.yaml
• Keep shared config in global
• Keep reusable config in component.commonConfig
• Keep instance differences minimal
• Avoid duplicating images/resources across instances
• Always preview before deploy:

helm template .