Case study

Elasticsearch Migration for a Video Commerce Platform

How SquareShift migrated a global video commerce platform's Elasticsearch system indices to Elastic Cloud using a dual-write cutover.

Book a session
3,500+Watchers updated through automated migration
4Fleet policies migrated and aligned
Selective migrationOnly system indices moved, skipping 52TB of historical data

A global video commerce company operating across live TV, e-commerce, and mobile platforms.

Its Elasticsearch-powered observability stack handles the log ingestion, monitoring, and alerting that keep its digital storefronts performing and available.

Impact

Watchers updated through automated migration. Fleet policies migrated and aligned. Only system indices moved, skipping 52TB of historical data.

Key services
ClCloud Modernization
PePlatform & Software Engineering
Industry

Media & Entertainment

Key technologies / platforms

Elasticsearch · Elastic Cloud · Logstash · Kibana · Elastic Fleet · SAML · Index Lifecycle Management (ILM)

The engagement

How SquareShift delivered it.

The challenge

The client, a global video commerce company spanning live TV, e-commerce, and mobile platforms, ran its Elasticsearch-powered observability stack on-prem — supporting the log ingestion, monitoring, and alerting that keep its digital storefronts running. A full migration would have meant moving 52 terabytes of historical data the platform didn’t need live.

Early attempts hit snapshot credential mismatches, ILM misconfiguration under the platform’s high ingestion volume, and API key and SAML authentication issues that were blocking dashboard access.

What we delivered

SquareShift migrated only the system indices the platform actually needed — Watcher, Kibana, and related indices — using dual writes through Logstash and Fleet, so historical data stayed in place and out of scope.

The team fixed the snapshot and SAML/API key issues, tuned ILM for the hot and frozen tiers under high ingestion, restored dashboard access, and migrated all four Fleet policies with matching index templates. Over 3,500 watchers were updated through automation rather than by hand.

The payoff

The client’s observability stack now runs its essential system indices on Elastic Cloud, with monitoring, alerting, and dashboards intact and Fleet policies aligned. Skipping the 52 terabytes of historical data kept the migration focused and avoided a costly, unnecessary data transfer.

You don't have to move fifty-two terabytes of history to migrate an observability stack — dual-writing the system indices your platform actually needs live is the cheaper, faster cutover.

Cloud Modernization Practice Lead, SquareShift