Can the the run time added to the phase report to be shown altogether with metric and log values? EG.: /CoreOS/m2crypto/Regression/upstream-tester/Test result: PASS metric: 0 Time: 10m 52s Log: /tmp/tmp.IjsZNh4536 DMesg: /tmp/dmesg.log For better comparision in performance and comparative testing. And for avoiding using time command within the scripts.
The comment you've mocked is a report_result output, it has nothing to do with rhtslib. I generally like the idea - we can add timestamps to the journal, and put the information in the rlJournalPrintText output. But I would treat it as informative as best - I don't think it is of any use as a performance metric.
General idea was to compare rebased packages in the comparative_workflow run, when you will be able see immediately the run time of both version and compare between each other.
It it true that emitting a warn when encountering some significant time raise in rcw could probably do no harm...
Created attachment 329084 [details] patch - start time and end time of phases stored in journal: <phase endtime="2009-01-15 11:32:32" name="fazicka" result="PASS" score="0" starttime="2009-01-15 11:32:30" type="FAIL"> - duration printed as a part of phase summary in rlJournalPrintText: :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: :: [ LOG ] :: fazicka :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: :: [ PASS ] :: Running 'echo abc' :: [ LOG ] :: Duration: 2s :: [ LOG ] :: Assertions: 1 good, 0 bad :: [ PASS ] :: RESULT: fazicka - I also took the liberty of replacing datetime module with time in journalling.py to make it work on RHEL3.
in main tree http://git.fedorahosted.org/git/beaker.git?p=beaker.git;a=commit;h=c36da67c05ed9ac88be7e8bd0f93839e46ec3625
I have just investigated some results and foud this info useful. VERIFIED.