I built bonsai.io/websolr.com while working at Twitter in 2010. I left in 2012, keeping a promise to a co-founder that I'd leave if we hit $1M ARR.

There were definitely a couple months of pager/ops overload where I hit 100hr workweeks and multiple hundreds of pagers per week. Hypervisor bugs meant that some tasks were hard to automate, and that automation would stochastically fail, so I stayed up til 4am watching the logs and correcting issues, then went back into the office at 10am.

Basically, it was a reasonable hedging strategy against either company's failure, but given both succeeded, I lost a good chunk of what I would have made if I was "all in" on either. Plus, it was about as miserable as it sounds, though I learned a ton.