← All case studiesLogistics · Elastic consulting

Case study

Storage Cost Cut 60% for a Logistics and Transportation Company

How SquareShift moved a logistics platform's Elasticsearch cluster to Elastic Cloud in staged phases, cutting storage cost 60% and finishing the full migration in 3 hours of downtime.

Book a session →
Illustration for the case study: Storage Cost Cut 60% for a Logistics and Transportation Company
7terabytes moved without a single all-at-once cutover
3hours of downtime for the full migration
1million shipments a year the platform supports

Client

A logistics and transportation company that moves more than a million shipments a year and runs technology-driven supply-chain operations.

Industry

Logistics

Services

PePlatform & Software EngineeringObObservability & Reliability

Stack

Elasticsearch · Elastic Cloud · Kibana · Frozen-tier storage (ILM)

How did SquareShift move the cluster to Elastic Cloud in staged phases?

01

The challenge

The client is a logistics and transportation company moving more than a million shipments a year on technology-driven supply-chain operations. Its Elasticsearch was running an old version that caused scale, cost, and performance problems, with more than 7 terabytes of data at stake and uptime that had to be protected.

02

What we delivered

SquareShift moved the cluster to Elastic Cloud in stages, moving data and traffic across in phases instead of a single cutover. Frozen-tier storage took over for cold, rarely-accessed data, cutting storage cost without touching what the team could still query day to day.

03

The outcome

The full migration cost the client just 3 hours of downtime, and frozen-tier storage cut storage cost by 60%. The client's own team now runs and monitors the platform on Kibana dashboards SquareShift rebuilt, after being trained to operate the new stack directly. Staged moves to Elastic Cloud are part of our Elasticsearch migration services.

“
A version jump this large rarely fails on the upgrade itself — it fails on the cutover. Moving data in stages, with uptime as the constraint rather than the version number, is what kept this migration to 3 hours of downtime.

Platform & Software Engineering Practice Lead, SquareShift

Related work