Right now the automatic system tests save a few strategic log files after each run in order to make it easier to debug errors after the fact. Whilst a good start, there are certainly more things we can do to improve upon this:
a) Perhaps our services should be configured to run with debug logging turned on?
b) We could consider saving system log files, and not just ours.
c) To reduce clutter, we could consider only saving log files on failures.