My latest project is full-on AWS orchestrated with Serverless (https://www.serverless.com/). Cognito, API Gateway, Lambdas, SQS, SNS, DynamoDB. Bit of a brick wall learning curve at first, but now that I'm past that, it is going shockingly smoothly.
Node for lambda language, httpapi / REST on the Api Gateway.
Very likely to use it again, with one possible change - perhaps using Auth0 or something else for the authentication. Cognito is far and away the weakest part of that stack.
My latest project is full-on AWS orchestrated with Serverless (https://www.serverless.com/). Cognito, API Gateway, Lambdas, SQS, SNS, DynamoDB. Bit of a brick wall learning curve at first, but now that I'm past that, it is going shockingly smoothly.
Node for lambda language, httpapi / REST on the Api Gateway.
Very likely to use it again, with one possible change - perhaps using Auth0 or something else for the authentication. Cognito is far and away the weakest part of that stack.
I used to use the Serverless stack a lot and I've found it to be really nice for a small team but a lot of our new projects are build on firebase.
All have pros and cons but I've found Firebase the fastest to move and adapt along the way.
Python / Django + Graphene, Graphql. For database: PostgreSQL from the get-go. Redis sometimes when I am using distributed workers (Celery).
Node or Go, PostgreSQL, REST.
That said — don't listen to what anyone here says. Use the stack in which you're most comfortable and productive.
Java (Spring Boot)
Aurora / ES
REST