“生产化” Lambda 和 API 网关

Question

I have recently been teaching myself how to build APIs using NodeJS + Express. I am still very much a novice but was really excited when they worked brilliantly in my dev environment. However, I struggled to know what and how to configure my server and codebase for production. There is lots of info out there to use things like Helmet and put NGINX in front as a proxy but some of those things are beyond me at this stage.

So I started to look at 'serverless' as a possible option for deploying them and have migrated my data to an AWS RDS PostgreSQL and setup a Lambda function that takes a query parameter and queries the database. I have also managed to configure AWS API Gateway that expects an API key to trigger the Lambda function.

My RDS and Lambda functions are restriction to a VPC, I am using an API Key and API Gateway defaults to HTTPS.

So I was after any advice or help on what other things I now need to consider as most of the traditional production challenges have been removed by using "serverless".

Answer 1

Here are some other things to think about:

1. Use stages and stage variables to set up a 'pipeline' in the application. In API Gateway, the stages are meant to represent this workflow, so you can have a beta stage and a prod stage. You can use stage variables to set up the Lambda function endpoints in the same way .
2. If you aren't already using Usage Plans , that is a good way to monitor usage by API Key and also set up rate limiting and quotas per API Key.
3. API Gateway and Lambda both publish CloudWatch metrics and logs, so you can monitor those and set up Alarms on the metrics as well.