Being reluctant to readily abandon "tools and ideas that worked for them in the past" in favor of adopting HN-buzzword-compliant technologies on flimsy pretext being presented by earnest people who nonetheless often have limited experience with the thing they're proffering and more often zero awareness of the similar fads that came before it is really, really far away from, "I am god. I've seen it all. I reject it all!"

Cost (dev, support, maintenance, etc.) conscious managers are going to ask tough questions like, "Why would we use Node.js and MongoDB to build a financial transaction processing switch?" and "Why would we deploy a Hadoop cluster to do offline analysis of only 4GB-5GB of transient data per day?" and "Why would we front Elasticsearch with Kafka when our write-load is well below what a small ES cluster can handle and the data is non-critical?" and "Why would we deploy Mesos for infrastructure of <20 machines that has no dynamic multi-machine scheduling needs and never will?"

There's an almost never-ending barrage of stuff like that I contend with. Which is fine. It's part of my job, and despite hard-earned cynicism... myself and my teams have still managed to build all manner of things with all manner of relatively esoteric tech.

Erlang, Rust, Elixir, Scala, Cassandra, Riak, HBase, Kafka, Storm, Spark, Ansible, Cobbler, Packer, Consul, Vault, LING, Rumprun.

No stranger here to including feedback and ideas to determine the best tools for the job that needs doing and their associated costs (dev, support, maintenance, etc.). I'm also no stranger to the flights of fancy that beset engineers who end up inventing reasons or overstate their problem space in attempts to justify using new shiny things.

Take it easy mate. Cheers!