Amazon Neptune isn't that great - discussions regarding vertical scaling

One thing that I love about Dgraph (there are many) is that it is open-source and easy to run. If I want to play around on my machine or boot up a micro-service architecture via docker compose or minikube locally I can. Its flexibility that I prefer to have as a professional engineer in team/platform settings too, not just as lone-contractor or hobbyist.

For example at SSENSE we leverage Kubernetes but some pieces of infrastructure (actually quite a few) still live in AWS. Bringing up new environments of our Kubernetes resources is easy be it deployed or just local. But the AWS parts at least in our experience are encumbered by issues ranging from just slowness of setup, to permissions issues, Region/Zone issues, dealing with the complexity of another toolchain apart from Kubernetes (e.g. terraform), needing coordination with the Ops team which can lead to occasional hard blocks, and so on. All to say that having a database that is flexible to run where you want is awesome IMO.

It's unclear if Neptune supports ACID transactions. They mention "you can use relationships to process financial and purchase transactions in near real time", which could be interpreted either way, and have nothing on the features page.

Still, IMO their design is outdated for something built in 2017.

Manish gives a fairly interesting run down on vertical scaling on Amazon Neptune. I'd like to hear some thoughts