Simplify detection of broken runs

[Zmicier Zaleznicenka]

To easier detect broken runs and discard them, the following is needed.

1. Each test run should have a field specifying version of tested software.
2. There should be an option to show time-series for all runs with given version and compare them with one of the runs with the same version.
3. Setting baseline/target runs should be easily done in bulk from UI.