Helm Chart Deployment

Helm is the recommended way to deploy Code Search on Kubernetes. It provides a simple, reproducible deployment with sensible defaults and easy customization.

Prerequisites

• Kubernetes cluster (1.25+)
• Helm 3.8+
• kubectl configured for your cluster

Quick Start

1. Add the Helm repository

   helm repo add code-search https://aanogueira.github.io/code-search
   helm repo update

2. Create a values file

   values.yaml

   config:
     database:
       url: "postgres://user:pass@postgres:5432/codesearch?sslmode=disable"
     redis:
       url: "redis://redis:6379"
   
   ingress:
     enabled: true
     host: code-search.example.com

3. Install the chart

   helm install code-search code-search/code-search \
     --namespace code-search \
     --create-namespace \
     -f values.yaml

4. Verify the installation

   kubectl get pods -n code-search
   helm status code-search -n code-search

Chart Values Reference

Global Settings

# Full values.yaml reference
nameOverride: ""
fullnameOverride: ""

# Image settings (applies to all components)
image:
  registry: ghcr.io
  pullPolicy: IfNotPresent
  tag: "latest" # Overridden per component if needed

imagePullSecrets: []

Configuration

config:
  # Database connection (PostgreSQL or MySQL)
  database:
    url: "" # Required: PostgreSQL or MySQL connection string
    maxOpenConns: 25
    maxIdleConns: 5
    connMaxLifetime: "5m"

  # Redis connection
  redis:
    url: "" # Required: Redis connection string

  # Zoekt settings
  zoekt:
    indexDir: "/data/index"
    reposDir: "/data/repos"

  # Indexer settings
  indexer:
    concurrency: 2
    cloneTimeout: "10m"

  # Scheduler settings
  scheduler:
    enabled: true
    pollInterval: "6h"
    checkInterval: "5m"

  # Repository settings
  repos:
    readonly: false

  # Connection settings
  connections:
    readonly: false

# Mount config from an existing secret
existingSecret: ""

API Server

api:
  replicaCount: 2

  image:
    repository: techquestsdev/code-search-api
    tag: "" # Defaults to chart appVersion

  service:
    type: ClusterIP
    port: 8080

  resources:
    requests:
      memory: "512Mi"
      cpu: "250m"
    limits:
      memory: "2Gi"
      cpu: "2000m"

  autoscaling:
    enabled: false
    minReplicas: 2
    maxReplicas: 10
    targetCPUUtilizationPercentage: 80

  nodeSelector: {}
  tolerations: []
  affinity: {}

Indexer Worker

indexer:
  replicaCount: 2

  image:
    repository: techquestsdev/code-search-indexer
    tag: ""

  # Extra environment variables
  # NOTE: CS_ZOEKT_URL is automatically set to http://localhost:6070
  # since Zoekt runs as a sidecar in the same pod
  env: {}

  resources:
    requests:
      memory: "1Gi"
      cpu: "500m"
    limits:
      memory: "4Gi"
      cpu: "2000m"

  nodeSelector: {}
  tolerations: []
  affinity: {}

Zoekt Search Server

zoekt:
  replicaCount: 1

  image:
    repository: techquestsdev/code-search-zoekt
    tag: ""

  service:
    type: ClusterIP
    port: 6070

  resources:
    requests:
      memory: "1Gi"
      cpu: "500m"
    limits:
      memory: "4Gi"
      cpu: "2000m"

Web UI

web:
  replicaCount: 2

  image:
    repository: techquestsdev/code-search-web
    tag: ""

  service:
    type: ClusterIP
    port: 3000

  resources:
    requests:
      memory: "256Mi"
      cpu: "100m"
    limits:
      memory: "512Mi"
      cpu: "500m"

Ingress

ingress:
  enabled: true
  className: nginx

  annotations:
    {}
    # kubernetes.io/tls-acme: "true"
    # nginx.ingress.kubernetes.io/proxy-body-size: "100m"

  host: code-search.example.com

  tls:
    enabled: true
    secretName: code-search-tls
    # Or use cert-manager:
    # clusterIssuer: letsencrypt-prod

Persistence

persistence:
  # Index storage
  index:
    enabled: true
    storageClassName: "" # Uses default if empty
    accessMode: ReadWriteMany
    size: 50Gi
    existingClaim: ""

  # Repository storage
  repos:
    enabled: true
    storageClassName: ""
    accessMode: ReadWriteMany
    size: 100Gi
    existingClaim: ""

Built-in PostgreSQL (Optional)

postgresql:
  enabled: false # Set to true to deploy PostgreSQL in-cluster

  auth:
    postgresPassword: ""
    username: codesearch
    password: ""
    database: codesearch

  primary:
    persistence:
      enabled: true
      size: 20Gi

Built-in Redis (Optional)

redis:
  enabled: false # Set to true to deploy Redis in-cluster

  architecture: standalone

  auth:
    enabled: false

  master:
    persistence:
      enabled: true
      size: 5Gi

Common Configurations

With In-Cluster Database

values-with-db.yaml

postgresql:
  enabled: true
  auth:
    password: "secure-password"

redis:
  enabled: true

config:
  database:
    url: "postgres://codesearch:secure-password@code-search-postgresql:5432/codesearch?sslmode=disable"
  redis:
    url: "redis://code-search-redis-master:6379"

External Database

values-external-db.yaml

config:
  database:
    url: "postgres://user:[email protected]:5432/codesearch?sslmode=require"
  redis:
    url: "redis://redis.example.com:6379"

High Availability (Shared Storage)

Multiple indexer workers sharing the same storage for parallel indexing:

values-ha.yaml

api:
  replicaCount: 3
  autoscaling:
    enabled: true
    minReplicas: 3
    maxReplicas: 10
  fileBrowsing:
    enabled: true  # Enable file browsing from API

indexer:
  replicaCount: 4  # Multiple workers share the queue
  persistence:
    accessMode: ReadWriteMany  # Required for multiple workers
    size: 200Gi

web:
  replicaCount: 3

# No sharding needed - workers share storage
sharding:
  enabled: false

Note

ReadWriteMany (RWX) storage is required for multiple indexer workers. Compatible storage classes: NFS, CephFS, AWS EFS, Azure Files, GlusterFS.

Hash-Based Sharding (Extreme Scale)

For very large deployments without shared storage:

values-sharded.yaml

api:
  replicaCount: 3
  autoscaling:
    enabled: true
    minReplicas: 3
    maxReplicas: 10

# Sharding distributes repos by consistent hash
sharding:
  enabled: true
  replicas: 5  # Creates indexer-0 through indexer-4
  federatedAccess:
    enabled: true  # Enable file browsing and replace via proxy
    indexerAPIPort: 8081

web:
  replicaCount: 3

# Each shard has its own storage (no RWX needed)
indexer:
  persistence:
    accessMode: ReadWriteOnce
    size: 100Gi

Tip

With sharding enabled, each indexer processes only repositories assigned to its shard. File browsing and replace jobs work via federated proxy to the correct shard. See Sharding Configuration for details.

With TLS and cert-manager

values-tls.yaml

ingress:
  enabled: true
  className: nginx
  host: code-search.example.com
  annotations:
    cert-manager.io/cluster-issuer: letsencrypt-prod
  tls:
    enabled: true
    secretName: code-search-tls

Installation Examples

Basic Installation

helm install code-search code-search/code-search \
  --namespace code-search \
  --create-namespace \
  --set config.database.url="postgres://user:pass@postgres:5432/codesearch" \
  --set config.redis.url="redis://redis:6379"

With Values File

helm install code-search code-search/code-search \
  --namespace code-search \
  --create-namespace \
  -f values.yaml

With Existing Secret

# Create secret with database URL
kubectl create secret generic code-search-config \
  --from-literal=database-url="postgres://user:pass@host:5432/db" \
  --from-literal=redis-url="redis://host:6379" \
  -n code-search

# Install with existing secret
helm install code-search code-search/code-search \
  --namespace code-search \
  --set existingSecret=code-search-config

Upgrading

# Update repo
helm repo update

# Upgrade release
helm upgrade code-search code-search/code-search \
  --namespace code-search \
  -f values.yaml

Uninstalling

# Uninstall release
helm uninstall code-search -n code-search

# Delete PVCs (WARNING: data will be lost)
kubectl delete pvc -l app.kubernetes.io/instance=code-search -n code-search

Troubleshooting

Check Release Status

helm status code-search -n code-search
helm history code-search -n code-search

Get Generated Manifests

helm get manifest code-search -n code-search

Debug Installation

helm install code-search code-search/code-search \
  --namespace code-search \
  --dry-run --debug \
  -f values.yaml

Rollback

helm rollback code-search 1 -n code-search

Next Steps

• Configuration - Detailed configuration options
• Code Hosts - Connect to GitHub, GitLab, etc.