Finite amount of test beds. Finite amount of runs possible in a month.
Multiple users. Release Ops. Development engineers from multiple teams.
And the 'easy going' quality managers mandating a 'run' for every unrelated development crowds up the testbed.
No real control.
And when, the analysis happens out of 800 odd runs a month, 400 can be policed. But now 'policing' is a 'job' and needs head count. Do not want to 'police' ? Demand to build more such test beds.
Sounds like a familiar problem even your org has ?
When the cacophony reached me 2-3 weeks ago from all the sub-orgs, I looked to traditional Indian political way of fixing (because everyone seemed to have enough justification for their irrational numbers, and even funnier, everyone agreed that the load can come down by 50% at the aggregate level).
Monthly QUOTA of 'runs' for each sub-group and a simple tweak to tool to dis-allow anything more than that from the team. Last couple of weeks, I ve not seen any deliverable of any team slipping because of this test bed.
A loss..
1 year ago
No comments:
Post a Comment