caching regression: opt-out from the caching just introduced for precise performance debugging

[Rdataflow]

Is your enhancement related to a problem? Please describe.
#966 suggests how to reproduce the SPARQL queries to test performance of Stardog. However the newly implemented cache #955 interferes the reproducibility of the performance tests.

Describe the solution you'd like
A way to re-show the same chart in visualize relying on the same SPARQL query executed and reporting performance in debug mode i.e. by introducing a new URL param cache=false or the like for debugging purposes.

Describe alternatives you've considered
wait until the cache might be empty (10 min) and hoping no one else triggered the cache in the meantime.

Usecases and impact
this will help to precisely know and report the relevant queries issues to the db operator in the context of #966

Implementation will be straight forward.
see also developers comment #955 (comment)