Ask HN: What is your go to stack for REST API?

17 points by ooooak ↗ HN

21 comments

[ 29.0 ms ] story [ 2016 ms ] thread
cakePHP on EC2 + postgres on RDS is nice. toss in some Redis caching and feed it traffic from a load balancer. you can get a lot done with very little effort.
Using EC2 for PHP and RDS for postgres isn't cheaper with Heroku than with AWS ?
yeh, it is quite a bit cheaper. However, heroku is more aggressive in throttling the cheap stuff. I still use it for ci though. AWS simply offers more granular control and a more robust infrastructure. If you dig down into it, you'll see heroku is actually running on AWS.
Python + Django on AWS Beanstalk backed by Postgres on RDS.

Django REST Framework provides pretty much everything you need for a REST api.

I just rolled a DRF backed app into production. Any tips or helpful dependencies you’ve had the pleasure of using?
If serialisation becomes a bottleneck with the built-in DRF serialisers I suggest you check out Serpy.
Rails + Postgres + Heroku. I may be living a decade in the past, but all the bugs are worked out and the development speed is extremely good.
Same here. Ruby is a pleasure to work with and Rails fits my brain.

The stack does not feel outdated at all.

Golang with Postgres. Love the simplicity and power of Go along with a binary executable.
Drupal 8
Django Rest Framework, Postgresql and Redis. RQ for background jobs.
Where my node folks at?

Node + Express + Nginx + Pgsql on Elastic Beanstalk.

We’re too busy building graphql apps instead of REST apps.
What is your experience with elastic beanstalk like?
It's just a beefed up EC2. I don't have high traffic so cannot say how well the autoscaling works but apart from that, not much different than putting your app on EC2.
Do you use any ORM or just raw SQL queries?
I use Sequelize as my ORM. There are debates around using/not using ORMs but personally I feel weird writing raw SQL in any language except a SQL terminal.
.NET Core + Web API

Postgres

Nginx

Digital Ocean

API gateway + Lambda functions (in Python) + DynamoDB if I need to store data.

Pricing is completely linear with usage (and very low/free for low traffic) and I don’t have to worry about uptime or scalability.