[Feature] optionally collect profiling data

[m-ildefons]

The current implementation of the support bundle contains more or less only the cluster state and logs. This is insufficient to debug some types of faults like unexpected high memory consumption or performance problems. These problems may also only occur in long-running clusters or at customer sites, making them hard to reproduce in development/test setups.
To aide debugging such problems, the support bundle should have an additional optional capability to collect profiling data from certain pods similar to the script here: https://github.com/rancherlabs/support-kb/wiki/One-Liners#rancher-profile-collection-script