kubectl-external-forward
kubectl-external-forward copied to clipboard
Controller with Envoy dynamic configuration
Problems to solve
- We need to wait for an Envoy pod (about ~10 sec)
- Cluster admin cannot enforce a pod configuration such as a node affinity or image pull secret
- Eventually garbage pod is remaining after termination
- It consumes much memory when many users run pods (10MiB per user)
Solution
Run a Kubernetes controller and shared Envoy pod.
TODO
- Investigate https://www.envoyproxy.io/docs/envoy/latest/start/quick-start/configuration-dynamic-filesystem