Skip to main content
The Getting Started flow shows that ParadeDB can search, retrieve vectors, rank, filter, and aggregate from inside Postgres. A production rollout is mostly about choosing the right topology, indexing the right fields, and giving Postgres enough room to do the work.

​
Production Checklist

  1. Choose the deployment model that matches your primary database.
  2. Load representative data, not only the small mock_items table.
  3. Design indexes around real query patterns.
  4. Tune index creation, reads, and writes with production-sized data.
  5. Confirm query plans with EXPLAIN.
  6. Decide how upgrades, backups, and high availability will work.

​
Production Readiness

ParadeDB Community, our open-source edition, has been downloaded millions of times. It is durable and crash-safe: changes to ParadeDB indexes are write-ahead logged and support crash recovery and point-in-time recovery. See Guarantees for the full details. ParadeDB Enterprise adds commercial licensing, support, and physical replication for ParadeDB indexes. Physical replication is what enables high availability, failover, and read replicas that can serve ParadeDB queries.

​
Deployment

Deployment Overview

Compare extension, self-hosted, logical replication, cloud platform, BYOC, and Enterprise options.

Install as an Extension

Add ParadeDB to a Postgres instance you already control.

Logical Replication

Replicate from managed Postgres when the primary cannot run ParadeDB.

High Availability

Plan read replicas, failover behavior, and production availability.

​
Index Design

Start from the queries your application actually runs. Add fields to the ParadeDB index when they need to participate in full-text search, vector retrieval, filtering, ORDER BY, GROUP BY, or aggregations.

Create an Index

Index table columns, tokenizer casts, and framework-native schema examples.

Choosing a Key Field

Pick the unique row identifier that ParadeDB stores in the index.

Vector Fields

Index embeddings alongside text and filters for nearest-neighbor and hybrid search.

Columnar Fields

Make scalar and tokenizer-backed fields available through columnar storage.

Performance Tuning

Tune index creation, read queries, and write-heavy workloads.

​
Operational Readiness

Before cutting over traffic, verify the important paths:
  • Run representative queries with EXPLAIN and confirm ParadeDB is used.
  • Test writes and updates against indexed tables.
  • Check backup, restore, and upgrade procedures.
  • Set monitoring around slow queries, index size, and segment growth.
When something does not look right, compare the query plan with the relevant query reference page and the Read Throughput notes.