Configuration Overview

Code Search can be configured using a YAML configuration file, environment variables, or a combination of both.

Configuration Methods

YAML Configuration File

Create a config.yaml file:

server:
  addr: ":8080"

database:
  url: "postgres://user:pass@localhost:5432/codesearch?sslmode=disable"

redis:
  addr: "localhost:6379"

zoekt:
  url: "http://localhost:6070"
  index_path: "./data/index"

The configuration file is searched in these locations (in order):

1. Path specified by CS_CONFIG_FILE environment variable
2. ./config.yaml (current directory)
3. ./config/config.yaml
4. /etc/code-search/config.yaml
5. $HOME/.code-search/config.yaml

Environment Variables

All configuration options can be set via environment variables with the CS_ prefix:

export CS_SERVER_ADDR=":8080"
export CS_DATABASE_URL="postgres://user:pass@localhost:5432/codesearch"
export CS_REDIS_ADDR="localhost:6379"
export CS_ZOEKT_URL="http://localhost:6070"

Environment variable names follow this pattern:

• Prefix: CS_
• Nested keys: separated by _ (uppercase)
• Example: server.addr → CS_SERVER_ADDR
• Example: scheduler.poll_interval → CS_SCHEDULER_POLL_INTERVAL

Precedence

When the same option is set in multiple places, this is the order of precedence (highest to lowest):

1. Environment variables
2. Configuration file
3. Default values

Tip

Use environment variables for sensitive data like database passwords and API tokens. Use the config file for other settings.

Configuration Sections

Section	Description	Documentation
server	API server settings	Server Configuration
database	PostgreSQL/MySQL connection	Database Configuration
redis	Redis connection	Redis Configuration
zoekt	Search engine settings	Zoekt Configuration
indexer	Indexer worker settings	Indexer Configuration
scheduler	Automatic sync scheduler	Scheduler Configuration
repos	Repository settings	Repository Configuration
replace	Search/replace settings	Below
sharding	Horizontal scaling	Sharding Configuration
ui	UI display settings	Below
rate_limit	API rate limiting	Observability
metrics	Prometheus metrics	Observability
tracing	OpenTelemetry tracing	Observability
codehosts	Code host connections	Below
connections_readonly	Lock connections UI	Below
repos_readonly	Lock repos UI	Below
Secrets	File-based secrets	Secrets Management

Complete Configuration Example

Here’s a complete config.yaml with all options:

# Server configuration
server:
  addr: ":8080"           # Listen address (host:port)
  read_timeout: 15s       # HTTP read timeout
  write_timeout: 60s      # HTTP write timeout

# Database configuration
database:
  url: "postgres://codesearch:secret@localhost:5432/codesearch?sslmode=disable"
  max_open_conns: 25      # Maximum open connections
  max_idle_conns: 5       # Maximum idle connections
  conn_max_lifetime: 5m   # Connection maximum lifetime

# Redis configuration
redis:
  addr: "localhost:6379"  # Redis address (host:port)
  password: ""            # Redis password (if required)
  db: 0                   # Redis database number

# Zoekt search engine configuration
zoekt:
  url: "http://localhost:6070"  # Zoekt web server URL
  index_path: "./data/index"    # Path to search indexes
  shards: 0                     # Number of shards (0 = auto)

# Indexer configuration
indexer:
  concurrency: 2                # Concurrent indexing jobs
  index_path: "./data/index"    # Path to search indexes
  repos_path: "./data/repos"    # Path to cloned repos
  reindex_interval: 1h          # Reindex interval
  zoekt_bin: "zoekt-git-index"  # Zoekt indexer binary
  ctags_bin: "ctags"            # Path to universal-ctags binary
  require_ctags: true           # If true, ctags failures will fail indexing

# Repository storage
repos:
  base_path: "./data/repos"     # Base path for repo clones

# Scheduler configuration
scheduler:
  enabled: true                 # Enable automatic scheduling
  poll_interval: 6h             # Default time between syncs
  check_interval: 5m            # How often to check for stale repos
  stale_threshold: 24h          # Max time before repo is stale
  max_concurrent_checks: 5      # Parallel git fetch checks

# Replace configuration
replace:
  concurrency: 3                # Parallel repo processing
  clone_timeout: 10m            # Git clone timeout
  push_timeout: 5m              # Git push timeout
  max_file_size: 10485760       # Max file size (10MB)

# Sharding configuration (horizontal scaling)
sharding:
  enabled: false                # Enable hash-based sharding
  total_shards: 1               # Number of indexer shards
  indexer_api_port: 8081        # HTTP API port for federated access
  indexer_service: "code-search-indexer-headless"  # Headless service name
  federated_access: false       # Enable file browsing/replace via proxy

# UI settings
ui:
  hide_readonly_banner: false   # Hide read-only mode banner
  hide_file_navigator: false    # Hide browse links in search results
  disable_browse_api: false     # Disable browse API endpoints

# Rate limiting
rate_limit:
  enabled: false                # Enable rate limiting
  requests_per_second: 10       # Requests per second per IP
  burst_size: 20                # Maximum burst size

# Prometheus metrics
metrics:
  enabled: true                 # Enable /metrics endpoint
  path: "/metrics"              # Metrics path

# OpenTelemetry tracing
tracing:
  enabled: false                # Enable tracing
  service_name: "code-search"   # Service name
  endpoint: "localhost:4317"    # OTLP endpoint
  protocol: "grpc"              # grpc or http
  sample_rate: 1.0              # Sampling rate

# Read-only modes
connections_readonly: false     # Lock connections to config-only
repos_readonly: false           # Lock repos to sync-only (no manual delete)

# Code hosts (optional - can also be configured via UI)
codehosts:
  github:
    type: github
    token: "$CS_GITHUB_TOKEN"   # Token or env var reference
    exclude_archived: true

  gitlab-internal:
    type: gitlab
    url: "https://gitlab.company.com"
    token: "$CS_GITLAB_TOKEN"
    exclude_archived: true

Read-Only Modes

Connections Read-Only

When connections_readonly: true:

• Connections can only be managed via the config file
• UI shows connections as read-only
• Add/Edit/Delete buttons are disabled
• Testing and syncing still work

This is useful for GitOps workflows where infrastructure is managed declaratively.

Repos Read-Only

When repos_readonly: true:

• Repositories can only be added/removed via sync
• Manual deletion via UI/API is disabled
• Excluding repos (soft delete) still works
• Excluded repos are cleaned from index and disk

Code Hosts via Config

You can define code hosts in the config file instead of (or in addition to) using the UI:

codehosts:
  # Use environment variable for token (recommended)
  github:
    type: github
    token: "$CS_GITHUB_TOKEN"
    exclude_archived: true

  # Self-hosted GitLab
  gitlab-internal:
    type: gitlab
    url: "https://gitlab.company.com"
    token: "$CS_GITLAB_TOKEN"
    exclude_archived: true

  # GitHub Enterprise
  ghe:
    type: github_enterprise
    url: "https://github.company.com"
    token: "$CS_GHE_TOKEN"
    exclude_archived: false

Tip

Use $ENV_VAR_NAME syntax for tokens to reference environment variables. This keeps secrets out of config files.

Config-defined code hosts take precedence over UI-created ones with the same name.

Validation

Code Search validates configuration on startup. If there are errors, the service will fail to start with a descriptive error message.

Common validation errors:

• Invalid URL formats
• Invalid duration formats (use Go duration strings like “5m”, “1h”)

Reloading Configuration

Configuration changes require a service restart. Hot-reloading is not currently supported.

# Docker Compose
docker compose restart api indexer

# Kubernetes
kubectl rollout restart deployment/code-search-api -n code-search
kubectl rollout restart deployment/code-search-indexer -n code-search

Next Steps

See detailed documentation for each configuration section:

• Server Configuration
• Database Configuration
• Redis Configuration
• Zoekt Configuration
• Indexer Configuration
• Scheduler Configuration
• Observability - Metrics, tracing, rate limiting
• Secrets Management
• Environment Variables