Elasticsearch Was Never a Database
This title could be clearer and more informative.Try out Clickbait Shieldfor free (5 uses left this month).
Elasticsearch excels as a search engine but fails as a primary database due to fundamental architectural limitations. It lacks atomic transactions beyond single documents, requires full reindexing for schema changes, has limited join support, and doesn't provide the durability guarantees expected from a system of record. Teams often start using Elasticsearch as a secondary index but gradually shift it to become their primary datastore, encountering consistency issues, operational complexity, and workarounds that negate the initial simplification. The article argues Elasticsearch should remain what it was designed to be—a search layer—while databases like Postgres handle transactional workloads.
