How to Set Up CI/CD with ParadeDB in PostgreSQL

How do I set up a CI/CD pipeline for ParadeDB in PostgreSQL?

ParadeDB CI/CD automates building, testing, and deploying a PostgreSQL instance running the ParadeDB extension across environments.

Welcome to the Galaxy, Guardian!

Oops! Something went wrong while submitting the form.

Description

Example H2

Example H3

Why use CI/CD for ParadeDB?

Automated pipelines catch schema issues early, guarantee extension compatibility, and ship new search features to prod with one click.

Which tools integrate best?

GitHub Actions, GitLab CI, and CircleCI all run ParadeDB in Docker images. Use Flyway or sqitch for migrations and pgTAP for tests.

How do I containerize ParadeDB?

# Dockerfile FROM paradedb/paradedb:latest ENV POSTGRES_PASSWORD=postgres COPY ./sql /docker-entrypoint-initdb.d

Place migration scripts in sql/. The image auto-executes them on start.

How do I write a GitHub Actions workflow?

# .github/workflows/ci.yml name: ParadeDB CI on: [push] jobs: test: runs-on: ubuntu-latest services: db: image: paradedb/paradedb:latest ports: ["5432:5432"] env: POSTGRES_PASSWORD: postgres steps: - uses: actions/checkout@v3 - name: Install psql run: sudo apt-get install -y postgresql-client - name: Run migrations run: psql -h localhost -U postgres -f sql/001_init.sql - name: Run pgTAP tests run: psql -h localhost -U postgres -f test/all_tests.sql

The pipeline builds, migrates, and runs unit tests on every push.

How do I promote to production?

Add a deploy job triggered on main merges that runs the same migrations against the prod database using a secret connection string.

What parameters matter?

Set PARADEDB_MAX_MEMORY and PARADEDB_MAX_WORKERS via ALTER SYSTEM SET in a migration file to keep prod and staging identical.

Best practice: use transactional migrations?

Wrap CREATE EXTENSION paradedb and config changes in a single transaction to avoid partial deploys.

Best practice: seed reference data

Load small product catalogs or test customers inside the pipeline so integration tests mimic production queries.

Why How to Set Up CI/CD with ParadeDB in PostgreSQL is important

How to Set Up CI/CD with ParadeDB in PostgreSQL Example Usage


-- Find top 5 products matching 'wireless headphones' using ParadeDB full-text search
SELECT id, name, price
FROM Products
WHERE to_tsvector('simple', name) @@ plainto_tsquery('wireless headphones')
ORDER BY ts_rank_cd(to_tsvector('simple', name), plainto_tsquery('wireless headphones')) DESC
LIMIT 5;

How to Set Up CI/CD with ParadeDB in PostgreSQL Syntax


-- Install ParadeDB extension
CREATE EXTENSION IF NOT EXISTS paradedb;

-- Tune ParadeDB settings (restart not required)
ALTER SYSTEM SET paradedb.max_memory = '2GB';
ALTER SYSTEM SET paradedb.max_workers = 8;

-- Example pipeline command sequence
psql -h $HOST -d $DB -U $USER -c "CREATE EXTENSION paradedb;"
psql -h $HOST -d $DB -U $USER -f sql/001_create_tables.sql
psql -h $HOST -d $DB -U $USER -f sql/002_create_indexes.sql

-- Ecommerce context: create a vector index for fast product search
CREATE INDEX idx_products_name_vector
  ON Products USING parade_vector (to_tsvector('simple', name));

Common Mistakes

Running CREATE EXTENSION paradedb in production only. Fix: add the statement to migration scripts so every environment runs identical DDL.
Skipping integration tests that check ParadeDB search functions. Fix: include pgTAP or pytest-pg fixtures to validate search accuracy during CI.