I hear you, it does seem like this year concerns have shifted drastically, it could just be collective nagging but I refuse to believe so. When it comes to the server-less concept I think it is stretching a bit too far, at least for now. There hasn’t been enough time for a true replacement to come forward and blow our expectations. Personally, servers are like pets to me, I’ll be hanging on to them for a very long time.
The over-tooling is becoming ever so present, lets just close our eyes and let it slide for now, it is a problem that will have to be gradually solved. A positive spin on it is that it fosters enough contributions and participation that there is a higher chance of something great emerging and make significant improvements.
Maybe your familiarity with the best practices and development of endpoints have made it second nature to you. For me, it is a clutter to have to think of an API and for my endpoints to be rigidly engineered to a dataset. If we need to make adjustment or even a new endpoint every time a major feature is under development I think of it as wasteful effort. GraphQL makes the data, all the data, readily available and CRUD operations dynamic in their composition. And that’s not even considering the a multiple endpoint call on every event or view. It is fantastic! If anything I would like you to give it a try.
If you would like external validation for the technology look no further that Github. They announce almost a year ago their release of the Github API v4 being built with GraphQl. Which to me makes a huge impact on the open-source community and other major companies.