What sorts of issues come up in a on-call alarm? When they come up, do you work to mitigate the problem forever, so that this particular alarm never happens aga

What sorts of issues come up in a on-call alarm? When they come up, do you work to mitigate the problem forever, so that this particular alarm never happens again? I am having a hard time imagining scenarios that need developers to be oncall. Is it a matter of pushing bad code to production?