Amazon Lambda, Serverless Framework, Python and RDS

I’ve been playing around with the Serverless Framework for setting up a simple API which fetches either one piece of data from a database or a list of things.

Setup

I decided on using Amazon RDS PostgreSQL as the database and figured that would be a pretty standard thing to do but within the Serverless framework it’s apparently not done that often. AWS has got a tutorial so it’s not completely impossible.

Challenges

  • Connection pooling / or not As the lambdas potentially could scale indefinitely and a standard RDS instance is just running like any other normal PostgreSQL database which allows a finite amount of connections. I decided to “fix it when it becomes a problem”.
  • How to install psycopg in a Python Lambda?
  • VPC access (or publicly available database)? Preferably via the Serverless Framework or at least via cloudformation

I have create a post for each of the latter two challenges:


Comments