LogicPlatform CRD Reference

The LogicPlatform CRD manages deployment of the data integration platform: Data Index (GraphQL query service) and Vector (log collection DaemonSet).

Overview

apiVersion: logic.kubesmarts.org/v1
kind: LogicPlatform
metadata:
  name: flow-platform
  namespace: data-index-system
spec:
  dataIndex:
    enabled: true
    persistence:
      postgresql:
        secretRef:
          name: postgresql-secret
        serviceRef:
          name: postgresql
          databaseName: logicflow
          databaseSchema: logicflow_dataindex
    vector:
      enabled: true
      watchNamespaces: ["workflows"]
      application:
        image: timberio/vector:0.54.0-distroless-libc

Spec Fields

dataIndex

Configures the Data Index service (GraphQL API) and supporting infrastructure.

dataIndex.enabled

Type Default

boolean

true

Enable or disable the Data Index Deployment and service. When disabled, no data collection occurs.

dataIndex.persistence

Specifies where Data Index stores workflow instance history.

persistence.postgresql

Required when Data Index is enabled. Provides PostgreSQL connection details for storing workflow events.

persistence.postgresql.secretRef

Reference to a Kubernetes Secret containing PostgreSQL credentials.

secretRef:
  name: postgresql-secret

The Secret must contain:

Key Value

POSTGRESQL_USER

PostgreSQL username

POSTGRESQL_PASSWORD

PostgreSQL password
persistence.postgresql.serviceRef

Reference to the PostgreSQL Kubernetes Service. The operator constructs the JDBC URL dynamically from the service endpoint. This field is required and cannot be omitted.

serviceRef:
  name: postgresql
  namespace: default  # Optional; defaults to LogicPlatform namespace
  databaseName: logicflow
  databaseSchema: logicflow_dataindex
Field Type Required?

name

string

Yes (Kubernetes Service name)

namespace

string

No (defaults to LogicPlatform namespace)

databaseName

string

Yes (must exist before Data Index starts)

databaseSchema

string

Yes (created by Data Index schema migrations)

port

int

No (defaults to 5432)

The database must exist in PostgreSQL before LogicPlatform is deployed. Schemas are created automatically: - LogicFlowRuntime uses the public schema (always exists in PostgreSQL) - Data Index creates logicflow_dataindex via Flyway migrations on first deployment

Ensure your PostgreSQL initialization includes the database creation step.

Use a single PostgreSQL instance for both runtimes and Data Index: - LogicFlowRuntime uses the public schema (default, no pre-creation needed) - Data Index uses logicflow_dataindex (created automatically by Flyway migrations)

This maintains logical isolation while avoiding manual schema creation.

dataIndex.vector

Configures the Vector DaemonSet for collecting and forwarding workflow logs.

dataIndex.vector.enabled

Type Default

boolean

false

Enable or disable Vector log collection. Requires dataIndex.enabled: true.

Vector is required to collect workflow events for Data Index. Without Vector, no structured logging events reach PostgreSQL.

dataIndex.vector.watchNamespaces

Type Default

[]string

Self-namespace (when empty)

Namespaces from which Vector collects workflow logs. When empty or omitted, defaults to watching only the LogicPlatform’s own namespace.

Examples:

  • [] or omitted — Watch only the self-namespace (default)

  • ["*"] — Watch all namespaces

  • ["workflows"] — Only the "workflows" namespace

Currently, only the first namespace in the list is monitored (WatchNamespaces[0]). Multi-namespace filtering support is not yet implemented.

By default, Vector watches only the LogicPlatform’s own namespace. For multi-namespace deployments, set watchNamespaces: ["*"] to collect from all namespaces, or deploy multiple LogicPlatform resources (one per namespace).

dataIndex.vector.debugEvents

Type Default

boolean

false

Enable debug logging of all workflow events to Vector’s stdout. This injects the DEBUG_EVENTS=true environment variable into the Vector container.

At 1000 events/sec, generates ~43 GB/day of logs.

Only enable for troubleshooting.

Example:

vector:
  enabled: true
  debugEvents: true
  application:
    image: timberio/vector:0.54.0

Disable after debugging by setting debugEvents: false.

dataIndex.vector.application

Type Required

ApplicationSpec

No

Configures the Vector DaemonSet deployment: image, replicas, resource requests/limits, and pod/container customization. When omitted, the admission webhook supplies default Vector image and resources.

Example:

vector:
  enabled: true
  watchNamespaces: ["*"]
  application:
    image: timberio/vector:0.54.0-distroless-libc
    replicas: 1
    resources:
      requests:
        memory: 128Mi
        cpu: 100m
      limits:
        memory: 256Mi
        cpu: 500m

Default resources:

Resource Request Limit

CPU

100m

500m

Memory

128Mi

256Mi

Adjust based on log volume: - Low volume (< 100 events/sec): Keep defaults - High volume (> 1000 events/sec): Increase to 512Mi/2000m CPU

Status Fields

Once deployed, the LogicPlatform status reflects the state of Data Index and Vector:

status:
  phase: Ready
  observedGeneration: 1
  dataIndex:
    service:
      ready: true
      deploymentRef:
        name: flow-platform-data-index
      serviceRef:
        name: flow-platform-data-index
      graphqlEndpoint: http://flow-platform-data-index.default.svc.cluster.local:8080/graphql
      replicas:
        desired: 1
        current: 1
        ready: 1
    persistence:
      valid: true
      secretExists: true
      serviceExists: true
      databaseConnected: true
    vector:
      ready: true
      daemonSetRef:
        name: flow-platform-vector
  conditions:
    - type: Ready
      status: "True"
      reason: Available
    - type: DataIndexDeploymentAvailable
      status: "True"
      reason: Available
    - type: DataIndexServiceReady
      status: "True"
      reason: Ready
    - type: DataIndexPersistenceReady
      status: "True"
      reason: Ready
    - type: VectorReady
      status: "True"
      reason: Ready

Key status fields:

Field Meaning

phase

"Ready" when Data Index and Vector are operational

dataIndex.service.ready

true when the Data Index deployment has at least one ready replica

dataIndex.service.graphqlEndpoint

GraphQL API URL. It uses the resolved external URL when Ingress is enabled; otherwise it uses the internal Service URL (for example, flow-platform-data-index.default.svc.cluster.local:80/graphql).

dataIndex.persistence.databaseConnected

true if PostgreSQL connection is healthy

dataIndex.vector.ready

true when the Vector DaemonSet is deployed and operational

conditions

Detailed status of each component (Deployment, Service, Database, Vector)

The graphqlEndpoint is an internal cluster address (using Kubernetes DNS names). To access Data Index from outside the cluster, use the external URL provided in status.dataIndex.service.url (when Ingress is enabled), or port-forward to the service: kubectl port-forward svc/flow-platform-data-index 8080:8080

Examples

Basic Setup (Watch Only Self-Namespace)

apiVersion: logic.kubesmarts.org/v1
kind: LogicPlatform
metadata:
  name: flow-platform
  namespace: workflows
spec:
  dataIndex:
    enabled: true
    persistence:
      postgresql:
        secretRef:
          name: postgresql-secret
        serviceRef:
          name: postgresql
          databaseName: logicflow
          databaseSchema: logicflow_dataindex
    vector:
      enabled: true
      application:
        image: timberio/vector:0.54.0-distroless-libc

This configuration deploys Data Index and Vector in the same namespace. Vector collects logs from workflows in that namespace only (default behavior). Data Index stores workflow history in logicflow_dataindex schema, while runtimes use the public schema.

Multi-Namespace Setup (Watch All)

apiVersion: logic.kubesmarts.org/v1
kind: LogicPlatform
metadata:
  name: flow-platform
  namespace: data-index-system
spec:
  dataIndex:
    enabled: true
    persistence:
      postgresql:
        secretRef:
          name: postgresql-secret
        serviceRef:
          name: postgresql
          databaseName: logicflow
          databaseSchema: logicflow_dataindex
    vector:
      enabled: true
      watchNamespaces: ["*"]
      application:
        image: timberio/vector:0.54.0-distroless-libc

This configuration deploys Data Index in a dedicated namespace and configures Vector to collect logs from all namespaces. Useful when workflow pods are distributed across multiple namespaces.

High-Volume Environment

For high-volume log ingestion (> 1000 events/sec), increase Vector resources:

apiVersion: logic.kubesmarts.org/v1
kind: LogicPlatform
metadata:
  name: flow-platform
  namespace: workflows
spec:
  dataIndex:
    enabled: true
    persistence:
      postgresql:
        secretRef:
          name: postgresql-secret
        serviceRef:
          name: postgresql
          databaseName: logicflow
          databaseSchema: logicflow_dataindex
    vector:
      enabled: true
      watchNamespaces: ["*"]
      application:
        image: timberio/vector:0.54.0-distroless-libc
        resources:
          requests:
            memory: 512Mi
            cpu: 500m
          limits:
            memory: 2Gi
            cpu: 2000m

Troubleshooting

Phase is "Pending"

Check component status:

kubectl describe logicplatform flow-platform

Look at conditions for which component is not ready: - DataIndexDeploymentAvailable - DataIndexServiceReady - DataIndexPersistenceReady - VectorReady

GraphQL Endpoint Empty

kubectl get svc flow-platform-data-index -o wide

Data Index Service may not be created yet. Wait for DataIndexServiceReady: True condition.

Vector Not Collecting Logs

kubectl logs -l app.kubernetes.io/name=flow-platform-vector

Common issues:

  • watchNamespaces filter doesn’t match workflow namespace

  • Vector pod pending (node tolerations issue)

  • PostgreSQL unreachable from Vector pod

See Workflow Instance History Troubleshooting for detailed diagnostics.

See Also