TanStack stack SSR in Kubernetes: Cache Management Across Multiple Pods

[btbenjamin]

Hello TanStack community,​

I'm currently deploying a React application using TanStack start with Server-Side Rendering (SSR) on a Kubernetes cluster. Given Kubernetes' nature of distributing requests across multiple pods, I'm interested in understanding how TanStack Router manages caching in this scenario:​

SSR Caching: Does TanStack Router's SSR implementation provide built-in mechanisms for caching rendered pages, or is this typically handled by external systems (e.g., CDNs, reverse proxies)?​

Cache Consistency: With multiple pods handling requests, how can one ensure cache consistency across these instances? Are there recommended strategies or best practices for synchronizing caches in a Kubernetes environment?​

State Management: How does TanStack Router handle state that needs to persist across different SSR requests, especially when those requests might hit different pods?​

Any insights, experiences, or resources on effectively managing SSR and caching with TanStack Router in a Kubernetes setup would be greatly appreciated.​

Thank you!

[schiller-manuel]

by default, nothing is cached by start itself.
you can opt into server function caching, see https://tanstack.com/start/latest/docs/framework/react/static-server-functions#what-are-static-server-functions