Case: I've got separate service that controls number of machines. It is designed in way that keeps control over number of instances. I mean, if one instance fails, it is restarted with the same IP. I see that there are ready-to-use solutions for AWS EC2, but the problem is, that I can't resign from this already implemented solution.
Question: Is there any solution that will allow me to administer containers in cluster, where number of machines, therefor workers isn't given in kubernetes config? Or do I have to rewrite everything from scratch?
Your question is pretty vague, but if you're looking for a solution that will auto-scale nodes (not Pods / or containers), there are options in AWS EC2 or GKE.
However, unless you have a really dynamic app, which has high fluctuations in CPU requirements, you will most likely be fine with a solution where you get a number of machines and scale the number of containers on that 'static' cluster of machines.
if you really need to scale nodes, look at:
https://kubernetes.io/docs/getting-started-guides/aws/#scaling-the-cluster
and
https://cloud.google.com/container-engine/docs/cluster-autoscaler