On the EWS layout test bots we are enabling core-dumps so we can get backtraces of the tests that crash.

And we have configured a cronjob to automatically clean cores older than 1 hour.

The issue is that sometimes when a patch tested causes too much crashes, due to the work the EWS for WPE and GTK work that repeats each failure 10 times then a lot of coredumps can be generated in a sort space of time. This may end causing the bot to run out of space.

This just happened here: https://ews-build.webkit.org/#/builders/86/builds/182

The server started with 41GB of free disk space (use 40%). See step  configuration
Then the step 'layout-test' caused 97 crashes
Then the step 'layout-tests-repeat-failures' tried to repeat each of those 97 crashes 11 times. So the Step would end generating 1067 coredumps :\
The server ran out of disk space and the run aborted due to this.

The next run on this server failed due to git issues (no disk space).

See: 

- https://ews-build.webkit.org/#/builders/86/builds/188
- https://ews-build.webkit.org/#/builders/86/builds/190
- https://ews-build.webkit.org/#/builders/86/builds/191
- https://ews-build.webkit.org/#/builders/86/builds/192
.. etc

Until an hour later the cronjob was fired and cleaned those coredumps the server was unusable.

So I think we need a way to configure the bots to just delete the cores after processing it and generating the backtrace.