I'm running my instance of shiny-server on kubernetes behind a load balancer. I've tried a variety of different shiny-server.conf configurations and am ultimately experiencing the same issue: my shiny server application runs fine on docker locally, but when I deploy my container to kubernetes, page loads produce a lot of 400 errors. The pattern in the errors seems to be something related to the "shared" directory resources and Uncaught Type/Reference Errors in the console. Sticky sessions are enabled.
I'm hoping someone might know a bit about how to diagnose and troubleshoot the problem. I've read the support article on running shiny server behind with a proxy, but I'm uncertain about whether the issues I'm having are due to websockets forwarding. I applied a patch to the envoy sidecar in my load balancer to allow connection upgrades, but this didn't have any effect (and my envoy configuration should upgrade to websockets automatically anyway). In addition to the 400 errors, I get the following console errors:
I'd be grateful for any guidance the community can provide!