Blocked productionizing this on results-2.html overhaul since we need it to merge the results from multiple system health benchmarks: https://github.com/catapult-project/catapult/issues/2345

We do not want to use results-1.html because all system health benchmarks are TBM2 based hence they produced a lot of metrics & most values are histogram object. results-1.html do not support visualizing lots of metrics & histogram well.

(marking the dependency on memory system health benchmark explicitly)

This is better be handled by pinpoint service instead. It's not practical to run all of the system health benchmark locally because:
1) It would take very long on a user machine. 
2) System health is a mixed of desktop & mobile benchmarks, so it's not that convenient for users to run all of them locally compared with using a trybot service.