Remote Reindexing Strategy for California Travel Startup

The challenge

The client, a California-based travel technology startup, ran an Elasticsearch deployment with 1,400 indices and 22 inconsistent field mappings — the kind of schema drift that builds up over years of ad hoc changes. Text-versus-keyword type mismatches and oversharding were both hurting performance and making the cluster harder to maintain.

More than 380 of those indices held little or no real data, adding overhead without adding value, and any fix had to account for both legacy and live data without extended downtime.

What we delivered

SquareShift unified the client’s mapping schemas and built custom index templates so every field type was consistent across the cluster. The team deleted the empty indices outright and consolidated the low-data ones rather than migrating them as-is.

Remote reindexing moved the cleaned-up data into Elastic Cloud, with ILM phases retuned and shard counts reduced through aliasing and targeted field exclusions.

The payoff

The client’s Elasticsearch footprint dropped from 1,400 indices with inconsistent mappings to a unified, right-sized cluster on Elastic Cloud, with more than 380 empty or low-data indices removed. Consistent field types and a leaner shard count leave the platform easier to maintain and faster to query.