PostgreSQL Configuration - CrewAI Platform Helm Chart

The internal PostgreSQL option is not intended for production use. It provides no high-availability, automated backups, or enterprise-grade reliability features. Always use an external managed PostgreSQL service (AWS RDS, Azure Database, Google Cloud SQL, etc.) for production deployments.

boolean

default:"false"

Enable or disable the built-in PostgreSQL subchart deployment.Values:

true - Deploy PostgreSQL as a StatefulSet in the cluster
false - Use external PostgreSQL database

When Enabled: The chart creates:

PostgreSQL StatefulSet with 1 replica
PersistentVolumeClaim for data storage
ClusterIP Service at <fullname>-postgres:5432
Automatic DB_HOST environment variable configuration

When Disabled: You must configure external database connection parameters:

envVars.DB_HOST
envVars.DB_PORT
envVars.DB_USER
envVars.POSTGRES_DB
envVars.POSTGRES_CABLE_DB
envVars.POSTGRES_OAUTH_DB
secrets.DB_PASSWORD

Pre-create databases: With an external database, you must manually create all required databases before deploying: primary, cable, OAuth, and Wharf (wharf by default). See the AWS or GCP guides for provider-specific instructions.

boolean

default:"false"

Danger Zone: When set to true, recreates the PostgreSQL data volume on next deployment, permanently deleting all data.

Use Cases:

Development environment reset
Testing clean installations
Never use in production

string

default:"crewai-postgres"

Override the full name used for PostgreSQL resources.Purpose: Controls the naming of PostgreSQL StatefulSet, Service, and other resources.Default: "crewai-postgres"Impact:

Service DNS name: <fullnameOverride>.namespace.svc.cluster.local
StatefulSet name: <fullnameOverride>
PVC name: data-<fullnameOverride>-0

When to Override:

Deploying multiple PostgreSQL instances in same namespace
Specific naming requirements
Avoiding name conflicts

Note: Changing this value after initial deployment may cause resource recreation.

string

default:""

Container registry hosting the PostgreSQL image.Default: "" (empty) - Automatically uses global.imageRegistry valueFallback Behavior:When postgres.image.host is empty or not set, the chart uses global.imageRegistry via the crewai-platform.postgresImageRegistry template helper. This allows you to configure the registry once at the global level.When imageNamePrefixOverride is Set:The image name is automatically simplified:

Original: proxy/crewai/dockerhub/library/postgres
With imageNamePrefixOverride: "crewai/" becomes: crewai/postgres

See global.imageNamePrefixOverride for details.Example:

global:
  imageRegistry: "registry.company.com"

postgres:
  enabled: true
  image:
    host: "" # Automatically uses registry.company.com
    name: "proxy/crewai/dockerhub/library/postgres"
    tag: "16"

string

default:"proxy/crewai/dockerhub/library/postgres"

PostgreSQL container image name.Default: "proxy/crewai/dockerhub/library/postgres" - Matches Replicated proxy path structurePath Transformation:When global.imageNamePrefixOverride is set, only the final component (postgres) is used with the override prefix.

string

default:"16"

PostgreSQL version to deploy.Supported Version: PostgreSQL 16Upgrade Consideration: PostgreSQL major version upgrades require data migration. Do not change this value without planning for database migration.

string

default:"IfNotPresent"

Image pull policy for PostgreSQL container.

string

default:""

Image pull secret for PostgreSQL image. If empty, uses the global image.pullSecret.

string

default:"crewai_plus_production"

Name of the primary PostgreSQL database created during initialization.

string

default:"crewai_plus_cable_production"

Name of the ActionCable database for WebSocket connections.Purpose: Rails ActionCable uses a separate database for pub/sub operations to isolate real-time communication from primary data operations.

string

default:"oauth_db"

Name of the Built-in Integrations service database for third-party integration token storage.Purpose: The Built-in Integrations service uses a separate database to store encrypted OAuth tokens for third-party integrations (Gmail, Google Calendar, Microsoft Outlook, etc.).Auto-Creation: When postgres.enabled: true and oauth.enabled: true, this database is automatically created during PostgreSQL initialization via an init script.Database Isolation: OAuth tokens are stored separately from the main application data for security and isolation.Default: "oauth_db"Related Configuration:

Built-in Integrations service: See CrewAI Built-in Integrations Reference
OAuth secrets: See Secrets Reference - Built-in Integrations Secrets

string

default:"wharf"

Name of the Wharf OTLP trace collector database for distributed tracing data storage.Purpose: The Wharf service uses a separate database to store OpenTelemetry traces and spans for application observability.Auto-Creation: When postgres.enabled: true and wharf.enabled: true, this database is automatically created during PostgreSQL initialization via an init script and before the Wharf service starts.Database Isolation: Trace data is stored separately from the main application data to isolate observability workloads and manage retention independently.Default: "wharf"Related Configuration:

Wharf service: See Wharf OTLP Trace Collector Reference
Wharf secrets: See Secrets Reference - Wharf Secrets

string

default:"cube"

Name of the Cube analytics database for cost tracking and execution analytics.Purpose: The Cube service uses a separate database to store analytics data, including provider usage metrics, cost tracking, and execution statistics.Auto-Creation: When postgres.enabled: true and cube.enabled: true, this database is automatically created during PostgreSQL initialization via an init script and before the Cube service starts.Database Isolation: Analytics data is stored separately from the main application data to isolate analytics workloads and manage retention independently.Default: "cube"Related Configuration:

Cube service: See Cube Analytics Reference
Cube secrets: See Secrets Reference - Cube Analytics Secrets

string

default:"10Gi"

Size of the persistent volume for PostgreSQL data.Sizing Guidelines:

Small deployment (< 100 users): 10-20Gi
Medium deployment (100-500 users): 50-100Gi
Large deployment (> 500 users): 100Gi+

Note: This value cannot be decreased after initial deployment. Plan for growth.

string

default:"keep"

PersistentVolume reclaim policy when the PVC is deleted.Valid Values:

"keep" - Retain volume data after PVC deletion (recommended)
"delete" - Delete volume data when PVC is deleted

string

default:""

Kubernetes StorageClass to use for the persistent volume.Common Values:

"" - Use default StorageClass
"gp3" - AWS EBS GP3 volumes
"managed-premium" - Azure Premium SSD
"standard-rwo" - GKE Standard persistent disk

Performance Consideration: Use SSD-backed storage classes for better database performance.

`postgres.resources.*`

Resource limits and requests for PostgreSQL container.

string

default:"1"

CPU limit for PostgreSQL pod.Sizing Guidelines:

Small deployment: "1"
Medium deployment: "2"
Large deployment: "4" or higher

string

default:"2Gi"

Memory limit for PostgreSQL pod.Sizing Guidelines:

Small deployment: "2Gi"
Medium deployment: "4Gi"
Large deployment: "8Gi" or higher

string

default:"500m"

Guaranteed CPU allocation for PostgreSQL pod.

string

default:"1Gi"

Guaranteed memory allocation for PostgreSQL pod.

object

default:"{}"

Kubernetes node selector for PostgreSQL pod placement.Example:

postgres:
  nodeSelector:
    node-role.kubernetes.io/database: "true"
    topology.kubernetes.io/zone: us-west-2a

Use Cases:

Pin database to specific nodes with fast local storage
Ensure database runs in specific availability zones
Separate database workload from application workload

​postgres.resources.*

`postgres.resources.*`