Case study
30 Terabytes Moved in Five Minutes of Downtime for an AgriTech Platform
How SquareShift migrated an AgriTech platform's 30-terabyte Elasticsearch cluster to Elastic Cloud with under five minutes of downtime for thousands of live users.
Book a session →
Client
An AgriTech company that delivers real-time agricultural insights and analytics, running search and monitoring for thousands of users in production.
Industry
Agriculture
Services
Stack
Elasticsearch · Elastic Cloud · Kibana · Google Cloud Pub/Sub · Index Lifecycle Management (ILM)
How SquareShift delivered it.
The challenge
The client's on-prem Elasticsearch cluster held 30 terabytes of data and served thousands of users in production. The system could not pause for the move — search and monitoring had to keep running the entire time.
Restoring that much data also risked overloading the busiest nodes on the new cluster under real load.
What we delivered
SquareShift wrote every update to both the old and new systems at once, through dual Google Pub/Sub topics, so the cutover never had to stop the platform. Roles, alerts, and agent configuration moved over through Kibana APIs and snapshots.
When JVM spikes appeared on the new cluster, the team resolved them by adjusting index allocation and tuning ILM policies. For a live move of your own cluster, see our Elasticsearch migration services.
The outcome
The move to Elastic Cloud took under five minutes of downtime, carrying over all 30 terabytes of data and more than 60,000 daily search hits. Daily operations continued on the new cluster without a gap.
A live migration succeeds or fails on the write path, not the data copy — writing to both systems at once is what let this cutover measure its downtime in minutes.
Platform & Software Engineering Practice Lead, SquareShift
