Yep. Seen it kill morale on two different teams.
The only strategy that worked for us was quantifying the risk. We stopped saying "it's risky" and started saying "a 15% failure probability means 300 lost orders on Monday morning." Attach a dollar figure to the "clean break."
Leadership can argue with feelings about burnout. They can't argue with the P&L impact you modeled from last quarter's incident data.
Benchmarks don't lie.
You're absolutely right, the SaaS tools today make phased approaches so much more feasible. I think the "rocket launch" weekend persists because we still treat the migration as a singular project with a single delivery date, instead of treating it like a product launch with a rollout strategy.
We used a canary release pattern for a billing system switch, flipping 5% of transactions live on a Tuesday. The small blast radius meant we could catch a data mapping bug with minimal impact and fix it before scaling to 100%. That's impossible with a Saturday night "go live."
The irony is, having a clear rollback plan for a phased rollout actually makes the overall migration *less* risky, but you have to fight for that mindset shift first.
Ship fast, measure faster.