parallel tests cancel is too drastic when one test failed

two concequences:
potential parallel tests may not complete successfully.
unable to capture logs of post_fail_hook of parallel tests to analyse

eg on rawhide 20181219 https://openqa.stg.fedoraproject.org/tests/overview?build=Fedora-Rawhide-20181219.n.1&groupid=3&distri=fedora&version=Rawhide

support_server  parallel failed
https://openqa.stg.fedoraproject.org/tests/442507
install_repository_nfs_variation    parallel failed
https://openqa.stg.fedoraproject.org/tests/442508
install_repository_nfs_graphical    parallel failed
https://openqa.stg.fedoraproject.org/tests/442510
install_updates_nfs failed
https://openqa.stg.fedoraproject.org/tests/442511
that one failed and only one to report logs 

The problem is new with updates of openQA;
this was not the case of last f29 tests (3 months ago)
https://openqa.stg.fedoraproject.org/tests/overview?distri=fedora&version=29&build=Fedora-29-20181024.1&groupid=3

Note that if I try to run locally the same f29 20181024.1 compose with new openQA,
then I experienced the same "parallel failed" state as above.

Is there a way to disable this cancel action, to help to capture data on parallel tests ?