Legacy platforms often begin as tightly coupled monoliths running on virtual machines. While workable at small scale, this model breaks down under real production traffic, operational growth, and security demands.

This talk presents a real-world case study of re-architecting a BI and AI analytics platform from a VM-based monolith into Kubernetes-native microservices serving millions of users. It covers a 13 years journey from on-prem bare metal and vertical scaling failures to containerization and a production-grade Kubernetes platform across cloud and on-prem environments.

The session focuses on practical challenges such as resource contention, downtime-driven scaling, observability gaps, security audit failures, configuration drift, and manual operations, and explains how these drove decisions around service decomposition, GitOps, defense-in-depth security, and service mesh adoption.