On-call burnout: strategies for sustainable incident management

on-call burnout is a serious problem on our team of eight engineers. we're getting 5-10 pages a week, and at least 2-3 of those are outside of business hours. we've already lost two engineers to burnout in the last year, and it's clear we can't sustain this. we've identified that about 85% of our alerts are non-actionable noise, which is a huge source of frustration. we're working on reducing that noise, improving runbooks, and investing more in self-healing systems. what strategies have you found most effective for creating a sustainable on-call rotation? beyond alert tuning, what else has made a real difference in preventing burnout for your teams?