Case study
Digital Commerce Firm Migrates to Elastic Cloud 8.3
How SquareShift migrated 2.4TB of sales-analytics data from Graylog to Elastic Cloud 8.3 for a digital commerce platform in 8 weeks, keeping storefronts live throughout.
An international digital commerce agency powering more than 45 global fashion brands.
It runs their e-commerce operations at scale, with real-time analytics and storefront reliability as the core promise.
Analytics data migrated across 9,000 indices. Billion documents migrated with live data capture throughout. Weeks from legacy Graylog to Elastic Cloud 8.3.
Key servicesHow SquareShift delivered it.
The challenge
A digital commerce agency powering more than 45 global fashion brands ran its sales analytics on Graylog over Elasticsearch 2.x — a stack that had become outdated and unstable. With no in-house Elastic expertise, the client needed to migrate to Elastic Cloud 8.3 in time for peak holiday-season traffic, without disrupting live storefronts.
The scale made the migration itself the hard part: 2.4TB of analytics data across 9,000 indices and 2.2 billion documents, on a tight, sales-event-driven timeline.
What we delivered
SquareShift designed a multi-step upgrade path — 2.x to 5.x to 7.x to 8.3 — rather than attempting a single risky jump. AWS Load Balancer and AWS SQS buffered and ingested GELF logs as JSON throughout the migration, keeping live data flowing while historical data was staged separately.
The team architected a hybrid migration strategy: historical and live data were separated, staged, and streamed using AWS SQS and custom ingestion scripts, keeping storefronts running throughout the switchover.
The payoff
The client completed the full migration — 2.4TB across 9,000 indices and 2.2 billion documents — in 8 weeks, moving from an unstable legacy stack to Elastic Cloud 8.3 in time for peak season.
Storefronts stayed live throughout the cutover, and the client’s LiveOps team was trained post-migration to operate the new platform directly, rather than depending on SquareShift for day-to-day operations.
A peak-season migration only works if you stop treating historical and live data the same way — stage the archive, stream the feed, and the cutover stops being a gamble.
Cloud Modernization Practice Lead, SquareShift
Where this work sits