Baseline ECK scale tests

[charith-elastic]

Part of #357

The goal is to avoid scale testing Kubernetes itself as the operator will always be bound by the limits of the underlying Kubernetes distribution. Instead, the plan is to observe the behaviour of a typical operator deployment when it is under pressure and gain an understanding of the bottlenecks and breaking points.

These baseline tests will run the operator using a reasonable set of resource requests and limits and observe the following.

• Number of resource instances (different combinations of reasonable ES, KB and APM deployments) that can be created using a given amount of RAM and CPU.
• Behaviour and resource usage of the operator when instances are randomly mutated/destroyed.