Case study
22 Always-On Nodes Cut to Two for a Global Fashion Retailer
How SquareShift moved a global fashion retailer's Elasticsearch cluster to Elastic Cloud with dynamic scaling, replacing a 22-node year-round footprint with a 2-node daily baseline.
Book a session →
Client
A global fashion and lifestyle retailer running multiple high-profile brands on an on-premises Elasticsearch cluster sized for peak season and kept over-provisioned year-round.
Industry
Retail
Services
Stack
Elasticsearch · Elastic Cloud · AWS · Microsoft · Index Lifecycle Management (ILM)
How SquareShift delivered it.
The challenge
The retailer's cluster was sized for its busiest shopping days — 22 nodes at peak — but ran that size every day of the year, paying for that footprint even in the slowest months. Its legacy, Microsoft-based application layer meant any migration had to work without touching existing integration code.
What we delivered
SquareShift moved the workloads to Elastic Cloud with full compatibility for the existing architecture, so the application layer needed no code changes. Secure network routing was set up between AWS and Elastic Cloud, and Index Lifecycle Management policies now scale the cluster to match actual demand.
The outcome
The cluster now runs a 2-node baseline day to day and scales up only when seasonal demand requires it, instead of staying at its 22-node peak size all year round. Moving workloads onto Elastic Cloud is covered by our Elastic migrations work.
Peak-season sizing kept year-round is just a subscription to your busiest day, paid every day. Dynamic scaling isn't a cost feature bolted on afterward — it has to be how the cluster is built from the start.
Platform & Software Engineering Practice Lead, SquareShift
