For performance regressions, delay adding the regressor needinfo after the first needinfo has been cleared

[arai-a]

bugbot shouldn't set needinfo again after it's cleared.
There's no point of setting it again and it's just annoying.

https://bugzilla.mozilla.org/show_bug.cgi?id=1959968#c2
https://bugzilla.mozilla.org/show_bug.cgi?id=1959085#c5

[marco-c]

Note the needinfo from the bot is completely independent from the initial needinfo, but I understand it can feel redundant given the performance regression needinfo is asking the same thing bugbot is asking.

In the first case, the needinfo was added soon after the previous one was cleared. For performance regressions where we know the needinfo is to the author of the regressor, we could implement a delay.
In the second case, the needinfo was added after 6 days, so I think it is acceptable given we want to encourage responding quickly to a regression (even if the response is just "not important, S4" or "wontfix" or "wrong regressor, clear the regressor field").

[suhaibmujahid]

we could implement a delay.

We could delay setting a needinfo if the regressor's author has cleared a needinfo after the "Regressed by" field was set. This would apply to the first case, not the second.