Then it either shouldn't be an alert (and instead part of some kind of summary report or some such) or the devs need to take on call. It is an exercise in frustration for everyone to route the page to ops just to make ops call dev; that means dev still has to have an oncall rotation, they might as well just take the page directly.
The unintended consequence of forcing alerts down ops' throat is them gradually caring less about pages, because there's a very good chance that each one is unactionable. I've worked places that do this, I've seen it happen first-hand more than once.
It starts with frustration and ops being less helpful to devs, and ends in a jaded acceptance where ops people start telling each other "just close it and see if it happens a second or third time, that alert never means anything". At that point, the system may as well not emit the errors anyways because no one is looking at the alerts anyways.
The unintended consequence of forcing alerts down ops' throat is them gradually caring less about pages, because there's a very good chance that each one is unactionable. I've worked places that do this, I've seen it happen first-hand more than once.
It starts with frustration and ops being less helpful to devs, and ends in a jaded acceptance where ops people start telling each other "just close it and see if it happens a second or third time, that alert never means anything". At that point, the system may as well not emit the errors anyways because no one is looking at the alerts anyways.