Companies of one, what is your tech stack in 2020?

Graphile Starter-based stack (Node, PostGraphile, Next.js, Graphile Worker) running on Heroku with Amazon RDS Postgres. Virtually no server maintenance needed: just push the code to GitHub, check it works on staging, press the promote to production button, job done. (I’m the Graphile maintainer.)

boring is key.

linux vm/dedicated (bare metal) server depending on the compute power I need

docker. plain docker. no kubernetes django htmx (the successor to intercoolerjs) postgresql traefik pycharm as my IDE of choice Scaleway, Hetzner, or DigitalOcean for hosting.

However, I now have some enterprise customers who want SSO. I will need to figure out how Keycloak fits in.

No remote storage like s3 No queues/streaming like kafka or rabbitmq (I use postgres for that)

javascript everywhere: client browser, nodejs for backend, ejs for template engine. except for storage, use sql (no mongodb).

try to check new stuffs, eg deno, nestjs, react, graphql, typescript, oauth2, etc. not worth to include them if it is not necessary, cost vs benefit.

prefer to use standard spec, use extra libraries as few as possible, e.g css variables vs SaaS, websocket vs socketio.

Also avoid npm complexity, except for express js. definitely, no webpack and their friends.

for server admin: systemd for process manager, nginx for reverse proxy/load balancing/cache.

Why? to preserve my brain power and focus. learn as few tech stack as possible, but master them. need to focus on product market fit.

Front-End: TypeScript, React, Bootstrap, Gulp and NPM.

Why: It's the most developed eco-system for front-end development. I also like TypeScript.

Back-End: Rust, Docker, Hashicorp (Terraform, Consul, Nomad, Vault)

Why: Because I use Rust now :) and the backend is an executable running inside a lightweight docker container.

Other: Github, DigitalOcean, Cloudflare

Why: The most developed eco-system again for Git and self-hosted containers. I hate Cloudflare but still didn't have time to explore alternatives.

Dev: Archlinux, i3, NeoVim, Alacritty

Why: By far the most productive development environment. I'm glad I moved from Windows to macOS and then from macOS to Linux.

While not for profit, I run a small community DB website for a podcast I like.

Front: Vanilla JS (just the littlest bit), Jinja2 templates

Back: Python + Flask, served via Gunicorn

Data: SQLite (mostly reads)

Server: Wrapped in a docker image and deployed on a $5/month VPS with a few other small websites I run. Small and easy to set up development or deployment.

what: python3 django, vanilla js, bootstrap, postgres, docker, nginx, aws

why: it's secure and I know it very well.

impact: yes it allowed me to move quickly, fix bugs efficiently, and keep a large project organized.

I'm solo developer

What:

Node.js/AWS lambda/DynamoDB/Kinesis for backend

Pytorch/k8s/AWS fargate for Deep learning api

Clingo/Prolog for symbolic AI

React.js/flexbox/sass for frontend

SwiftUI for iOS

Why: each tool is optimal for its use case

Impact: move fast and break things)

scala for backend, scala.js for frontend (react), postgres

its my favorite language and i can express myself very quickly and in a very safe way. Postgres because i like strict schema that ensures data consistency.

scala allows me to work quickly so that is one impact, its also a typesafe language so that is another obvious benefit.

what: ruby on rails, postgres, htmx front end, deployed to heroku

why: keep things as simple as possible

impact: Yes, to an extent, it allowed me to keep my head around the whole system. I don't think the platform you pick is a huge part of your success or failure though.

Django, Vue, Postgres @ aws

I base my language choice on $/line, which means I use a lot of php

nextjs + react, postgraphile + postgres, apollo client, rds, heroku

Frontend : Svelte

Backend : Java (Quarkus)

DB : PostgreSQL

Deployed on a VM via docker.

Not that fancy, but they just work.