> And not consider whether there are better solutions such as checking we're using the right algos or cache-aligning our accesses or using a probabilistic filte

> And not consider whether there are better solutions such as checking we're using the right algos or cache-aligning our accesses or using a probabilistic filter before hitting the DB or.. It's mostly for io-intensive workloads. And for nicer debugging experience compared to callback hell. Nothing to do with NoSQL, or right algos, or cache-aligning when you just need to do 10M http requests/db-inserts/rpcs as fast/efficient/nicer-dev-experience as possible.