I chuckled at this: > "Do you see the impact you created for thousands of us without any warning or explanation? We are not your test subjects," said an angry sysadmin. "We are running professional services for multi million dollar programs. Do you understand how many hours of resources were wasted by your 'experiment'?" Did you, dear sysadmin, pay anything for Google Chrome? No? Are you in any contractual relation t…
Yeah. > We are running professional services for multi million dollar programs. Do you understand how many hours of resources were wasted by your 'experiment'? -- Angry sysadmin Well, do they pay Google for testing in their environment? Chrome is a free-to-use product. Their rollout strategy is good. Not that they experiment in prod - the flag was in beta for 5 months. And then they turned that on in prod for %1 user…
I think you've misunderstood what they did
> the flag was in beta for 5 months. And then they turned that on in prod for %1 users, still no reports.
It was in beta, had no reports, and then they turned it on for a subset of users, and created a flood of problems.
> Well what better could they do?
Release it in a new version that doesn't auto-enable for small subsets of people, but is enabled for everyone if they've deployed that version. It's how software releases used to work. Enterprise environments can then test before it gets rolled, and save this problem from occurring.