AgriTech Elasticsearch Migration to Elastic Cloud

The challenge

The client is an AgriTech platform whose search and monitoring systems support thousands of users in production. Moving 30 terabytes of Elasticsearch data from an on-prem cluster to Elastic Cloud had to happen without pausing live ingestion or breaking the access, alerting, and agent configuration the platform already depended on.

Restoring that much data also risked hot node saturation on the new cluster — a problem that would only show up under real load, after the cutover was already underway.

What we delivered

SquareShift ran the migration through dual Google Pub/Sub topics, capturing live writes in parallel so the cutover never had to pause ingestion. Kibana APIs and snapshots carried over every role, alert, and agent configuration intact.

On the new cluster, the team resolved JVM spikes by adjusting index allocation and tuning ILM policies, then validated the environment and handed over documentation and training before calling the migration complete.

The payoff

The cutover to Elastic Cloud took under five minutes of downtime, with all 30 terabytes and 60,000-plus daily search hits carried over intact. Configuration, alerting, and access controls moved with the data, so the platform’s day-to-day operations picked up on the new cluster without a gap.