← Graph

Default to Postgres with good indexes before reaching for Elasticsearch

takeaway 1 connections

Consensus recommendation from the panel, especially Stephen and Caio: before adopting Elasticsearch for a big-dataset filterable index endpoint, use Postgres (or SQLite) with proper indexes, load-test across environments, and find the Pareto-hot filter combinations. Only reach for a secondary index store if you actually outgrow the relational database. Don't fear joins — Postgres handles 20+ joins fine.

type
recommendation
takeaway Default to Postgres with good indexes before reaching for Elasticsearch
from_talk
Consensus response to the 20+ filter endpoint question.

Provenance

Created in
Performance Panel at wroclove.rb 2024 2026-04-17 23:20