Production Checklist

​

Database

• Use a production-ready Postgres instance: SQLite is great for development but Postgres is recommended for production workloads due to better concurrency handling and scalability.
• Configure connection pooling: Ensure your Postgres connection pool is appropriately sized for your worker concurrency.
• Set up backups: Regular database backups are critical since your workflow state lives in the database.

​

Workers

• Run at least one worker process: Deploy workers as separate long-running processes. They can run on the same servers as your application or separately.
• Concurrency: Start with concurrency: 10 per worker and tune based on your workload.
• High availability: Run multiple worker instances for redundancy. Workers are stateless and can be scaled horizontally.

​

Error Handling

• Review your step functions to ensure proper error handling
• Configure appropriate retry policies for your use case
• Set up alerting for persistently failing workflows

​

Debugging

• Use the dashboard to inspect workflow runs: npx @openworkflow/cli dashboard
• Review step attempts and errors for failed workflows
• Check database state for workflow runs that appear stuck

​

Namespaces (Optional)

const backend = await BackendPostgres.connect(postgresUrl, {
  namespaceId: "production",
});