Modernizing News Search for a Trading Firm

The challenge

A global proprietary trading firm ran two custom-built applications at the center of its news infrastructure: a Node.js ingestion service handling thousands of items a second, and a proprietary search interface built on top of it. Both were maintenance-heavy and hard to scale further.

The client needed to replace both with something built on standard components — without losing the throughput or the advanced search features high-frequency trading decisions depended on.

What we delivered

SquareShift replaced the Node.js ingestion app with a Logstash pipeline built on persistent queues and retry logic, engineered to sustain throughput of 2,000 news items a second. The team rebuilt the search experience in Kibana, replicating Lucene query support, highlighting, and filtering from the original proprietary interface.

The full solution — Logstash and Kibana — was deployed on Kubernetes, replacing bespoke infrastructure management with standard, container-native operations.

The payoff

The client now runs on production-grade Elastic Stack components instead of two custom-built applications, sustaining 2,000 news items a second through a fault-tolerant Logstash pipeline with role-based access control.

Legacy apps were decommissioned in favor of fully managed, Kubernetes-based services, cutting maintenance overhead while improving performance and giving the team standard tooling to build on going forward.