ClusterAPI Operator optimization and refactor

[alexander-demicev]

CAPI Operator consumes a lot of CPU and Memory while processing providers. We need to identify problematic spots and fix them.

• Optimize configMap cache usage in repeated reconciles without changes #133
• ✨ Enable profiling for operator pod kubernetes-sigs/cluster-api-operator#810
• Reduce CPU usage by optimizing large yaml processing

[Danil-Grigorev]

Blocked on reviews for kubernetes-sigs/cluster-api-operator#827

Both unit tests and e2e tests were fixed after another round of self review on the PR (typo happened due to long waiting time on review of the PR and subsequent context loss)