✨ [POC] Scale tests & Tracing

[sbueringer]

Some thoughts on Tracing:

• Instead of wrapping the reconcilers we should have another util func that we call at the beginning of each top-level Reconcile func (should do the same as traceReconciler.Reconcile at the moment)
• we should align flags to upstream k/k / component-base
• we should align the way we create the TracerProvider to k/k / component-base
• we have to decide for which funcs we open new spans (e.g. only the ones which are expected to take some time (e.g. because of network calls))
• we should instrument the client-go http client. It's already implemented for the client for the mgmt cluster. There were problems with the workload cluster client (random spans in the trace that took 10s)
• we should implement a dashboard which makes it easy to analyze trace metrics (first version is included in this PR)