The stack is now set up in one place. The S3 bucket isn't created on create if a bucket is already specified. Working on configurability for the developer. There is still the possibility of failing to deploy the stack if the iamRoleArn is set, and a deploymentBucket is specified
Deploy
This plugin (re)deploys the service to AWS.
How it works
Deploy starts by hooking into the deploy:setupProviderConfiguration lifecycle.
It fetches the basic CloudFormation template from lib/templates and replaces the necessary names and definitions
with the one it gets from the serverless.yml file.
Next up it deploys the CloudFormation template (which only includes the Serverless S3 deployment bucket) to AWS.
In the end it hooks into deploy:deploy lifecycle to update the previously created stack.
The resources section of the serverless.yml file is parsed and merged into the CloudFormation template.
This makes sure that custom resources the user has defined inside the serverless.yml file are added correctly.
Note: Empty, but defined Resources or Outputs sections are set to an empty object before being merged.
Next up it removes old service directories (with its files) in the services S3 bucket. After that it creates a new directory
with the current time as the directory name in S3 and uploads the services artifacts (e.g. the .zip file and the CloudFormation
file) in this directory. Furthermore it updates the stack with all the Resources which are defined in
serverless.service.resources.Resources (this also includes the custom provider resources).
The stack status is checked every 5 seconds with the help of the CloudFormation API. It will return a success message if
the stack status is CREATE_COMPLETE or UPDATE_COMPLETE (depends if you deploy your service for the first time or
redeploy it after making some changes).