Agreed, test issues showing up in blockers/critical is confusing. I guess, for product issues found through automated tests, we could use Issue type = Bug and component = Automation in addition to the actual component where the bug is found.
> -----Original Message----- > From: Prasanna Santhanam [mailto:[email protected]] > Sent: Sunday, July 28, 2013 5:23 PM > To: [email protected] > Subject: Re: New Components on JIRA > > On Wed, Jul 24, 2013 at 10:15:12AM +0000, Ram Ganesh wrote: > > > > > > > > Prasanna, > > > > > > > > How about - automation-product and automation-script components? > > > > Automation-product for all product bugs discovered by the > > > > automation engine and automation-script for all automation script > > > > issues? > > > > > > > > > > Right now - that distinction is not clear at least from the bug > > > reports. We're reusing the same report for both script and product > > > failure. So anything filed from an automated test failure should > > > just be automation and on further analysis if it is found to be > > > product failure, a clearer bug report would be necessary within the > > > right component of the product - api, network, systemvm etc. > > > > > > > Yes you are right. Component is not the right field. Maybe we could > > use Label. A JIRA report which reports product issues discovered out > > of automation will be very valuable. Maybe label is a field for that. > > > > So I noticed there is are a few issue types in JIRA - Bug, Improvement, Task > and > Test. We should use 'Test' for any thing related to automated tests. This > will help > the RMs filter out bugs that are not entirely product failures but may be > missing/failing tests. Right now having automated tests show up in blocker and > critical lists is a little confusing. > > I'm not really sure how we can track product issues caught by automated tests > other than simple labelling may be? Would be a good report to have to reveal > our coverage and strengthen the suites. > > The lifecycle of a bug as I see it is something like this > > Test fails on jenkins > \ > --> Test failure filed under issue type 'Test' > \ > --> If cloudstack issue, bug report filed with repro steps and logs > | > --> OR Test fixed appropriately, reviewed and pushed > > Similarly, for any bug filed, > > Bug found in product > \ > --> Identify if missing from tests and is automatable > \ > --> Add a 'Test' issue for addressing the missing test > \ > --> Push test to repo and add it to jenkins runs > > These may be [sic]'noble' goals, but it would be great if we collectively > start the > effort of identifying areas that can be covered using tests. > > -- > Prasanna., > > ------------------------ > Powered by BigRock.com
