Customer work · Retail and E-commerce
Lenskart partnered with BluePi to design a cloud-native enterprise analytics platform on Google Cloud, supporting more than 2 TB of daily ingestion, 6000+ MySQL tables, and near real-time CDC latency for retail operations.
Lenskart’s analytics spanned AWS DMS, Redshift, S3, MySQL, MongoDB, VM-based CRON jobs, custom APIs, and Azure-hosted Power BI. Manual ingestion, VM-based orchestration, and limited governance constrained scale. BluePi consolidated that estate into a governed, near real-time analytics platform on Google Cloud.
01 · Fragmented multi-cloud estate
Analytics workloads ran across AWS, on VMs, and in Azure-hosted Power BI. Manual ingestion, missing CI/CD and version control, limited observability, and little business ownership of transformation logic constrained scale and governance.
02 · Cloud-native platform on Google Cloud
Debezium on GKE handled enterprise-scale CDC, Pub/Sub carried event-driven ingestion, and BigQuery Raw, Silver, Gold, and Reporting layers with Dataform transformations supported a domain-oriented analytics architecture.
03 · Governed near real-time analytics
The platform supports more than 2 TB of daily ingestion, 6000+ MySQL tables, and sub-minute CDC latency, with Dataplex, Cloud DLP, column lineage, IAM, VPC Service Controls, and automated data quality monitoring in place.
BluePi combined Debezium on GKE for CDC, Pub/Sub ingestion, layered BigQuery storage, Dataform transformations, and Dataplex governance into one platform, then added Terraform, Cloud Build, CI/CD, and SmartMigrate acceleration to create an engineered operating model.
Multi-cloud and VM-based operations were consolidated into one governed Google Cloud platform before scale requirements were applied.
Terraform, Cloud Build, CI/CD, Cloud Monitoring, Cloud Logging, and alerting made the platform a repeatable, observable product instead of a one-off build.
Workload assessment, migration planning, SQL modernization strategy, inventory analysis, transformation lineage analysis, and architecture generation were accelerated so engineering effort focused on business-critical transformation.
Where this pattern fits
This pattern fits retail organizations whose analytics spans multiple clouds, VMs, and unmanaged integrations and whose reporting needs to support day-to-day operations across supply chain, warehouse, finance, marketing, CRM, logistics, and sales. A useful first step is one business domain with a measured ingestion, latency, and governance baseline.
Open a section to review the customer problem, implementation, business change, and architecture.
Lenskart had grown a complex analytics landscape across AWS DMS, Redshift, S3, MySQL, MongoDB, VM-based CRON jobs, custom API integrations, and Power BI hosted on Azure. The platform needed to consolidate multi-cloud analytics operations and replace legacy OLTP reporting patterns with a stronger OLAP foundation.
BluePi designed an end-to-end modern data platform on Google Cloud and an engineered operating model around it.
The platform gave Lenskart a governed, near real-time analytics foundation for retail operations and a base that supports AI and ML initiatives.
“BluePi demonstrated deep expertise in enterprise data platform modernization and cloud architecture. Their team did the majority of heavy lifting not just to lift and shift the data, but to rewrite legacy code and pipelines. They architected the new platform that now ingests more than 2 TB of daily data, has 6000+ operational tables, hundreds of integrations, and 250+ Power BI dashboards. Their expertise in the BigQuery ecosystem is commendable. The proposed architecture delivers near real-time data availability, enterprise-grade governance, and a scalable foundation for analytics across our business.”
