DATA MODERNIZATION
Global Financial Services
Scalable Event Architecture for Finance
Summary:
Our client, a Global Financial Services Leader, is driving an ambitious digital transformation program. The organization is moving toward a cell-based architecture; a modular, scalable framework designed to enhance resilience, efficiency, and innovation.
A core part of this transformation is the creation of a new Operational Data Store (ODS) capable of handling real-time event data while ensuring consistency, performance, and cost efficiency at scale. To achieve this, the client needed a modern event-processing backbone that could power materialized views and real-time insights across
the business.
Tech Stack:
- Database: MongoDB Atlas
- Cloud Platform:AWS (EKS, MSK, Glue Schema Registry)
- Backend Framework: Java Spring Boot
- Event Streaming: Apache Kafka, Kafka Streams, MSK Connect
- Infrastructure as Code: Terraform, Helm, GitHub Actions
- Monitoring & Observability: OpenTelemetry, Prometheus, Loki, Tempo, Grafana
- Testing: Python API and load testing suite
gravity9 partnered with a global financial services organization to modernize its fragmented, legacy data landscape and support a shift toward a resilient, scalable cell-based architecture. The team designed and delivered a fully functioning event-driven Event Sink built on MongoDB Atlas, Kafka, and AWS EKS. Working within strict access constraints and a shared AWS environment, gravity9 engineered a lightweight but powerful platform comprising an event generator, Kafka Streams processor, and an Event API, all deployed through automated Terraform and Helm pipelines. Full observability was implemented using the OpenTelemetry stack, Prometheus, Loki, Tempo, and Grafana, enabling real-time visibility across ingestion, transformation, and persistence stages.
The Proof of Concept validated that MongoDB Atlas could reliably serve as the Event Sink for large-scale, event-driven workloads within the client’s emerging architecture. The system met all performance benchmarks, including sustained throughput of 1,500 events per second, stable autoscaling, reliable failure recovery, and real-time event processing. With more than 600,000 documents processed at consistent latency, the PoC demonstrated the architecture’s resilience and scalability. This success has unlocked the next phase an MVP focused on production deployment and extending the event-sink model to additional use cases across the client’s digital ecosystem.
RELATED INSIGHTS
Talk to our Development team
Ready to scale without compromise? Partner with gravity9 to future-proof your platform. Talk to our team today.
Contact Us
"*" indicates required fields