add-pgvector-db

name	add-pgvector-db
description	Deploy pgvector to K8S and configure a Quarkus app for the connection to the DB

Add pgvector Database on Kubernetes

Description

Deploys a PostgreSQL 15 database with the pgvector extension on Kubernetes/OpenShift and configures the Quarkus application to connect to it. This enables vector similarity search for RAG, embeddings, and AI/ML workloads.

When to Use

• User wants to add a vector database to their Quarkus application
• User needs pgvector for RAG (Retrieval-Augmented Generation) or embedding storage
• User asks to set up PostgreSQL with vector support on Kubernetes

Steps

1. Apply Kubernetes Manifests

Apply the manifests in this skill directory to the target namespace:

# Set the target namespace (use current context namespace if not specified)
NAMESPACE=$(oc project -q 2>/dev/null || kubectl config view --minified -o jsonpath='{..namespace}')

# Apply the PVC, Secret, Deployment, and Service
kubectl apply -f .claude/skills/add-pgvector-k8s/pgvector-pvc.yaml -n $NAMESPACE
kubectl apply -f .claude/skills/add-pgvector-k8s/pgvector-secret.yaml -n $NAMESPACE
kubectl apply -f .claude/skills/add-pgvector-k8s/pgvector-deployment.yaml -n $NAMESPACE
kubectl apply -f .claude/skills/add-pgvector-k8s/pgvector-service.yaml -n $NAMESPACE

2. Wait for the Database to be Ready

kubectl rollout status deployment/pgvector -n $NAMESPACE --timeout=120s

3. Add Quarkus Dependencies (if not already present)

Add the following dependencies to pom.xml:

<dependency>
    <groupId>io.quarkus</groupId>
    <artifactId>quarkus-hibernate-orm-panache</artifactId>
</dependency>
<dependency>
    <groupId>io.quarkus</groupId>
    <artifactId>quarkus-jdbc-postgresql</artifactId>
</dependency>

4. Configure Quarkus Datasource

Add the following to src/main/resources/application.properties:

# PostgreSQL pgvector datasource
quarkus.datasource.db-kind=postgresql
quarkus.datasource.username=${POSTGRESQL_USER:quarkus}
quarkus.datasource.password=${POSTGRESQL_PASSWORD:quarkus}
quarkus.datasource.jdbc.url=jdbc:postgresql://${POSTGRESQL_HOST:pgvector}:${POSTGRESQL_PORT:5432}/${POSTGRESQL_DATABASE:quarkus}
quarkus.hibernate-orm.database.generation=update

5. Verify the Connection

# Port-forward to test locally
kubectl port-forward svc/pgvector 5432:5432 -n $NAMESPACE &

# Verify pgvector extension is installed
PGPASSWORD=quarkus psql -h localhost -U quarkus -d quarkus -c "SELECT extname FROM pg_extension WHERE extname = 'vector';"

Manifests Included

File	Resource	Purpose
pgvector-pvc.yaml	PersistentVolumeClaim	20Gi storage for database data
pgvector-secret.yaml	Secret	Database credentials (user/pass/dbname: quarkus)
pgvector-deployment.yaml	Deployment	PostgreSQL 15 with pgvector, lifecycle hook to enable extension
pgvector-service.yaml	Service	Exposes PostgreSQL on port 5432

Configuration Details

• Image: quay.io/rh-aiservices-bu/postgresql-15-pgvector-c9s:latest
• Port: 5432
• Default credentials: user=quarkus, password=quarkus, database=quarkus
• Storage: 20Gi PVC at /var/lib/pgsql/data
• Memory limit: 512Mi
• pgvector extension: Automatically created via postStart lifecycle hook

Important Notes

• Change the default credentials in pgvector-secret.yaml before deploying to production. Values are base64-encoded.
• The PVC uses the default StorageClass. Specify a storageClassName if needed.
• The postStart hook waits for PostgreSQL readiness before creating the vector extension.
• For production, consider increasing the PVC size and memory limits.