KubeCon + CloudNativeCon India 2026

In life-critical systems such as 911 dispatch, emergency call handling, and public safety platforms, the traditional “one cluster per application” model creates operational silos that limit resilience, scalability, and cost efficiency. This session describes how we challenged that model by building a Shared First Kubernetes platform that allows many sensitive services to coexist without compromising isolation or availability
We operate 25+ production Kubernetes clusters across multiple clouds. Each shared cluster runs 50–60 independent products and supports 2,000–4,000 microservices processing real-time emergency workloads where downtime can impact human lives.
The talk covers zero-touch workload isolation using a custom Kubernetes Admission Controller, per-tenant NGINX Ingress Controllers, automated DNS and certificate management and safe upgrade practices. We conclude by showing 50–60% infrastructure cost savings while improving reliability for mission-critical public safety systems.