Post by Florian Forster

In the world of cloud-native identity, performance isn't just a metric—it's a feature. When your application scales to millions of users, every millisecond of latency in the authentication flow directly impacts user experience and revenue. For B2B SaaS applications, where complexity grows with every new customer organization, this challenge is acute. As part of our 2026 Vision for Reliability at Scale, we are rolling out a flexible, high-performance caching architecture for ZITADEL. We are moving away from "black box" infrastructure. We now support three distinct caching connectors to match your specific deployment complexity: 🔹 Redis-Compatible: For high-scale, distributed Kubernetes deployments where cache consistency across pods is non-negotiable. 🔹 PostgreSQL: The stable default that leverages existing infrastructure to handle massive traffic loads (30k+ RPS) without added operational overhead. 🔹 In-Memory: For edge scenarios and rapid development. Strategic Impact: This architecture solves the "N-over-N" scaling bottleneck. By caching Instance & Organization contexts, we flatten the hierarchy. This allows application pods to scale linearly without punishing the database, effectively decoupling read performance from data complexity. Identity infrastructure should evolve alongside your deployment needs, not dictate them. Read the full blog: https://lnkd.in/gqaaQmTG #CloudNative #SystemDesign #Golang #IdentityManagement #SaaS #Redis