Jeff Dean pointed out in 'The Tail at Scale' that if a single request calls 100 backend services in parallel and each service has a P99 latency of 10ms, the probability of the entire request completing within 10ms is only about 36.6%.