Are you using Rancher in production?

Question

Kubernetes

Launched for testing on two nodes using containers. One node (etcd, controlplane, worker), the second agent (worker). Confuses a few points:
1. Is the control server not fault-tolerant?
2. What will happen if he dies (that's it)?
3. Didn't find the console on the web, do I need to fall through the exec of the container?
Advise what you use to deploy and manage k8s clusters? To a greater extent, it is the deployment and scaling of workers and masters that is of interest. Web faces and prettiness are optional

Answer 1

Are you using Rancher in production?

Yes.
Failsafe.
Nothing.
The metrics and Node Pools will simply fall off - when restarting, it may not pick up the state and it will be necessary to add / remove the handles of the "lost" nodes in the Node Pool.
This will not affect the working application particularly if you have more than 3 machines in the pool ... otherwise it can lie down.
The Kubectl of the corresponding classter is launched ... you can already knock from it anywhere.
We ourselves are sawing an analogue of the rancher on steroids, with chess and poetesses ...