sampique v0.7.0

SAMPIC

Deploy utility for AWS Serverless Application Model (SAM) based projects that significantly speeds up time to deploy updates when the only change is the code of Lambda function and the rest of the stack is the same.

It builds upon aws cloudformation package but will figure out if simply updating code of Lambda functions or updating stack parameters is enough or whether a full aws cloudformation deploy is required.

Install

Install this package globally.

npm install -g sampic

Generate a sample config file for your current git working branch:

cd <project_directory>
sampic init

usage

Make sure you have AWS CLI (version >= 1.11) and git installed and available in the shell environment when running sampic.

Make sure your current working branch is the branch to deploy and simply run the sampic command from your project base directory. Note that changes saved but not yet committed will be bundled in the deployment.

sampic [<options>] [<command>] [<command-options>]

The utility supports the following set of commands:

Manage npm dependencies

Deploy updates

Other commands

Configuration Setup

Create a .sampic directory at the base of your project and add a config.json file under it. CloudFormation templates packaged by the aws cloudformation package will also go under this directory. It is a good idea to add it to your .gitignore.

The configuration file declares deployment parameters for git branches you want to deploy. For example, the config file defines deployment instructions when your working branch is master or my-dev-branch:

{
  "master": {
    "profile": "default",
    "region": "us-east-1",
    "template": "app.yaml",
    "stackName": "my-staging-stack",
    "stackParameters": {
      "ParameterKey1": "ParameterValue1",
      "ParameterKey2": "ParameterValue2"
    },
    "s3Bucket": "lambdafns-staging",
    "capabilities": ["CAPABILITY_IAM"]
  },
  "my-dev-branch": {
    "profile": "default",
    "region": "us-east-1",
    "template": "app.yaml",
    "stackName": "my-dev-stack",
    "s3Bucket": "lambdafns-dev",
    "capabilities": ["CAPABILITY_IAM"]
  }
}

config parameters:

• profile (optional) refers to a named profile under ~/.aws/credentials. No need to define this if you'll use the default AWS profile
• region is the AWS region your CloudFormation stack is deployed in
• template is your SAM template file
• stackName is the name of the CloudFormation stack this branch should be deployed to
• stackParameters (optional) passed as template parameters when deploying the template to CloudFormation
• s3Bucket is the bucket where artifacts (lambda function code, external swagger files, ...) are uploaded to
• capabilities is used when running to the aws cloudformation deploy command. See AWS docs for more but usually, deploying SAM templates need at least CAPABILITY_IAM listed in there.

Application with multiple templates/stacks

If your application has more than one template deployed to different stacks, sampic supports that but the configuration is a bit different.

{
  "master": {
    "profile": "default",
    "region": "us-east-1",
    "stacks": {
      "myStackA": {
        "template": "stackA-template.yaml",
        "name": "stack-A"
      },
      "myStackB": {
        "template": "stackB-template.yaml",
        "name": "stack-A",
        "parameters": {
          "ParameterKey1": "ParameterValue1",
          "ParameterKey2": "ParameterValue2"
        }
      }
    },
    "s3Bucket": "lambdafns-staging",
    "capabilities": ["CAPABILITY_IAM"]
  }
}

Once you're config is set up with a <branch>.stacks option, commands parsing your CloudFormation template will list stack names and ask you which template you want to work with:

$ sampic deploy
Using config for current git branch: master
Which stack?
  (1) stackA-template.yaml => stack-A
  (2) stackB-template.yaml => stack-B
Specify stack number:  

Alternatively, you can use the stack key (<branch>.stacks.<key>) from CLI to skip the prompt:

$ sampic deploy --stack myStackA

What actually happens when you run deploy-local

The script goes through the following steps:

1. Gets current git working branch and looks for corresponding settings in .sampic/config.json
2. Based on the configured stack name, load the currently deployed template.
3. Package your template by running aws cloudformation package. The output template file is saved as .sampic/<STACK_NAME>-packaged-template.yaml
4. If no corresponding stack was found in CloudFormation, deploy the packaged template to create it. Otherwise, compare the new packaged template to the current stack template:
   • If the only difference is code updates for Lambda functions (stack parameters defined in the config haven't changed), simply update function code of those Lambdas.
   • If the packaged template is identical to the deployed stack template but stack parameters in the config are different, run aws cloudformation update-stack using the current template and specifying new parameter values.
   • If there are more changes, run aws cloudformation deploy to do a full stack update.
5. If all went well, rename .sampic/<STACK_NAME>-packaged-template.yaml as .sampic/<STACK_NAME>-deployed-template.yaml and we're done.