> ... find their approach much more flexible vs lambda@edge being possibly triggered four times at edge part of a request life cycle.
The lamda@edge function is executed only once but you decide when - before or after the cache and either on the request or response.