slsmanager 0.1.18

A basic tool that simplifies the deployment of AWS lambda functions

Serverless Manager

A basic Python Tool that simplifies the deployment of multiple AWS Lambda functions directly from an AWS CodeCommit Repository.

pip install slsmanager

Set Up

• Set up an IAM user on your AWS console

  • You require the following permissions:
  • Store the Access Key ID and the Secret Access Key
  • Generate the git credentials for the user. (Security Credentials -> HTTPS Git Credentials)

• Install serverless-manager on your local machine:

  • pip install slsmanager

Using the tool

On your Local Computer:

• Open a terminal window and navigate to the directory with the handler file.
• Run serverless manager on your terminal: slsmanager

1. Creating a CodeCommit Repository

• Follow the on screen instructions to Create a CodeCommit Repository and clone it to your directory using the Git Credentials.
• You can also do this from the AWS console.

2. Creating a CodeBuild Project

• Create a CodeBuild Project (OPTION 1: using slsmanager)
  • You can use slsmanager to create a CodeBuild Project.
  • You will need your AWS Access Key ID, your Secret Access Key, the default region and the URL to your CodeCommit Repository
• Create a CodeBuild Project (OPTION 2: using AWS console)
  • Skip these steps if you used slsmanager to create the project
  • Set the source provider to CodeCommit and the repository to the newly created CodeCommit repository
  • Pick the branch you will be pushing to
  • In buildspec, choose “Use a buildspec file” and leave the filename empty (Since the default is buildspec.yml
  • Create the project
  • Edit the environment variables and add the following variables:
  • The ENV_NAME_ variables refer to the different environments you would be deploying. For instance, I have 3 environments: dev, prod and uat. This allows us to deploy the lambda function at different stages (Creating a different function for each stage)

3. Creating the CodePipeline

• Serverless manager doesn't support the creation of a CodePipeline yet.
• Using the AWS console:
  • Source provider: AWS CodeCommit
  • Repo Name: Source repo
  • Build Provider: AWS CodeBuild
  • Project Name: Name of Codebuild Project
  • Skip deploy stage since we are using the Serverless Framework to deploy our functions
  • Create the Pipeline

4. Creating serverless.yml and buildspec.yml

If you have a file called handlerFun1.js, and it contains the function firstFun() which you would like to deploy.

• Follow the on-screen instructions:
  • Here, I created a service called "CodeCommitTest"
  • I set the AWS-region to us-east-2 (which is what my AWS console is set to).
  • The stage sets the serverless stage property.
  • The ENV_NAME is the last part of the ENV_NAME_ environment variables (which we set earlier). This will be reflected in the final name of your function on your console. This property allows us to create multiple deployed versions of the same lambda function source.
  • Module name follows the convention directory/filename.function
  • The function name is the name of the final lambda function (and is set in serverless.yml).
  • By default, the only file added to the current lambda function (That will show up on the console) is the handler file provided in the module name. If you would like to include other files in the final lambda function that is deployed, you may edit the newly generated serverless.yml and add the required files to the functions/your_function/package/patterns
  • You may also edit/add any other properties in serverless.yml
  • Finally, the tool generates a buildspec.yml file for your function.
    • If you want to deploy all lambda functions present in serverless.yml you can press 'a' in the final step. Otherwise, press 'n'.
  • Once you have completed all your steps, you will see 2 new files serverless.yml (Which lists all the properties required to deploy using the Serverless Framework and buildspec.yml (Which instructs CodeBuild to deploy the function(s) using the given files and properties.

Directly through AWS CodeBuild (> v0.1.7) :

• Alternatively, Serverless Manager can directly be used from AWS CodeBuild.
• Set up the required CodeCommit Repo, CodePipeline and the CodeBuild Project.
• Create a buildspec.yml file using the provided template.
• Create an environment variable SERVICE_NAME in the CodeBuild project and set its value to whatever you want the service to be called.
• Add this buildspec.yml file to the repository along with your handlers
• For now, the js handlers need to be called index.js and python handlers need to be called lambda_function.py. Place these handlers in sub-directories (UI, testing, backend).
• Once you commit/push these changes, CodeBuild will automatically run Serverless Manager and create a serverless.yml file (which will be removed after deployment).

Options

• There are a few config options you can use while running slsmanager

Add / Create

slsmanager --add/-a

• This allows you to skip the preliminary steps in the CLI (Create CodeCommit Repo/ Create CodeBuild Project) and jump directly to adding functions to serverless.yml and (later) buildspec.yml.

Buildspec

slsmanager --buildspec/-b

• This allows you to skip all preliminary steps in the CLI and jump directly to adding deploy commands to buildspec.yml.

NoCLI

slsmanager --nocli/-n

• REQUIRES -o to also be used
• This allows you to deploy the functions directly from codeCommit without needing to create a serverless.yml. Follow the steps above.

Options

slsmanager -n -o "service" "region"

• Options required for the noCLI method.