Request to collect/attach more logs for stress tests

Currently, after running the following stress tests:

* reboot, poweroff with pm_test
* s3, s4 with fwts

the output logs "pm_test.[reboot/poweroff].[repetitions].log" and "results.log" contains roughly the following:
- actions performed by pm_test or fwts
- fwts checks

However, these info are not enough for debugging when a pm_test cycle fails.

(e.g. for pm_tests. a failed test always ends with logs like "Failed to resume within [timeframe]", which is not very useful information.
We don't know whether the failure occurred during boot up or shutdown, what process was causing the failure, at which cycle did the test fail...etc.)

We still have to manually go through system logs and compare them with plainbox logs to figure out at what stage and for what reason the failure occurred.

---------------------------------

Therefore we would like the propose the following request:

1. Insert the output of "journalctl -x" that occurred during the test to the log, hopefully alongside the current existing pm_test/fwts logs in chronological order.

2. If the above method will result in a log file that is too large in size, create a separate file/tarball that contains the only the logs of the last cycle.

e.g. If a total of 3 cycles are run with command "pm_test reboot -r 3, then in the end this file/tarball should only contain the log for the 3rd cycle of test (with info also indicating that this is the 3rd cycle , this part would be bug #1576570).

This way we will always have a way to quickly check the logs of the last round, thus having a way to quickly check the logs of a failed cycle.