Skip to main content
Blnk is optimized out of the box for most workloads. This guide is for teams processing millions of records who want to squeeze out extra performance when searching directly from the DB.

​
Default indexes

These fields are already indexed and optimized:
You can sort and filter efficiently on any of these indexed fields without additional configuration.

​
Use Typesense for large-scale queries

For complex searches, filtering across millions of records, or when you need sub-millisecond response times, use Blnk’s Typesense integration instead of database queries. Typesense is optimized for:
  • Full-text search across any field
  • Faceted filtering with instant results
  • Sorting large datasets without custom indexes
See Search via Typesense docs for more.

​
Adding custom indexes

If you’re not using Typesense and frequently sort or filter by a specific field at scale, add an index.
1

Create single column indexes

Create indexes for frequently filtered or sorted fields:
2

Create composite indexes

For queries that filter and sort on multiple fields, composite indexes are more efficient:
3

Create metadata indexes

If you frequently filter on a specific metadata key, index that key directly:
The meta_data field already has a GIN index for general JSONB queries, but indexing specific keys provides better performance for frequent filters on those keys.
Use CONCURRENTLY when creating indexes on production databases to avoid locking tables. This allows the index to be built without blocking reads and writes.

​
Debugging slow queries

Use EXPLAIN ANALYZE to see how Postgres executes a query:
What to look for:
  • Seq Scan on large tables: strong signal you need an index
  • High actual time values: indicates slow query execution
  • Large row counts being filtered down late: suggests missing indexes
Example Output
In this example, a sequential scan is being performed, which suggests adding an index on status would improve performance.

​
Best practices

  1. Monitor query performance: Use EXPLAIN ANALYZE regularly to identify slow queries.
  2. Index strategically: Only add indexes for fields you filter or sort on frequently. Too many indexes can slow down writes.
  3. Use composite indexes: When filtering and sorting on multiple fields together, create composite indexes.
  4. Consider Typesense: For complex queries or very large datasets, Typesense often provides better performance than database indexes.
  5. Test in production-like environments: Index performance can vary based on data distribution and query patterns.

​
Need help?

We are very happy to help you make the most of Blnk, regardless of whether it is your first time or you are switching from another tool. To ask questions or discuss issues, please contact us or join our Discord community.