Mix serveless and manual operations in AWS console

Hi friends!

First of all, sorry if this question is off topics. Just point me where should write it instead and Ill do it.

So. I bought this liveVideo on manning to learn more about Serverless framework.
So far its super nice but at some point, for certain configurations (ex: api gateway usage plan, cognito)
The instructor do not do it in the serverless.yml but in the console directly.

Is is the right things to do? I get the point that it’s a training and he want to move fast to its next subject. But he did not mention at all if we should just stick to the serverless.yml or not. I feel we should but I would like to have you opinions too!

Hi !

I just realized a serverless project, and I initialized some base ressources directly in the console (VPC, RDS, Cognito, …). It works pretty well like that.

I think we can do better. The idea that I will test is to get one project to initialized these base ressource, and then other projects reference the base ressources of this stack. But I didn’t tried it yet.

Does it make sense ?

Having more than one serverless stacks and one to manage resources make sense to me so far. My conclusion is that Ill try to manage everything from the configuration file. If I have to redeploy my stack or update it, I thing it will be easier to manage.

if you are curious I did post the same question in these 2 subredits:

Well, doing things manually is the worst thing you can do in the Cloud. If you are learning, OK for the first time to get things maybe quickly setup, but after you see it working, destroy it and put all the resources inside your YAML file.

1 Like

I have multiple Cloud Formation (CF) stacks in the same AWS account, as well as manually creates resources. Some of the CF stacks are managed by ElasticBeanstalk, and some by Serverless. The Cloud Formation stacks all share the manually created resources, such as subnets, security groups, databases and internet/nat gateways. It works fine, and I so far have had no issue with Serverless or Elastic Beanstalk trying to modify the shared resources. I have considered putting the shared resources in another Cloud Formation stack, but it seems unnecessary right now.

So, do whatever makes sense to you, but make sure to have a clear definition of what should be managed by Serverless so that you don’t end up fighting it when manually modifying resources.

1 Like