setup.md

Setup

The included fabfile can faciliate setting up LambdaMLM.

1. Need pip, Virtualenv, and Fabric installed.
2. Clone this repo.
3. Copy config.example.py to config.py and edit/fill in the appropriate values. In particular, your S3 bucket name must be globally unique, not just unique within your account.
   Note: If you're planning to use multiple configurations, you can keep each configuration in a file named config.somename.py and pass somename as a parameter to the fab commands create_lambda:somename and update_lambda:somename. The specific somename configuration will be copied over config.py before the rest of the fab command runs.
4. In the directory, run fab setup_virtualenv to set up the virtual environment for LambdaMLM and install required dependencies.
5. Run fab create_lambda to create the S3 bucket, an IAM role under which LambdaMLM will run (with an appropriate policy), and the lambda function itself.
6. In SES, in the region defined as lambda_region in config.py:
   1. Verify all domains to be used for lists.
   2. Configure DKIM and SPF for domains.
   3. Create an Email Receiving rule that applies to all domains to be used for lists with two actions:
      1. S3: Store to the S3 bucket defined in your config.py with the incoming email prefix defined in config.py (example is incoming/).
      2. Lambda: Invoke the lambda function as an Event.

If you make any further changes to config.py or to the code, run fab update_lambda to update the Lambda function's code with your local code.

Technical Details

The fab create_lambda command:

• does a quick check of your config.py file

• creates the S3 bucket in the specified region.

• creates an IAM role with the name defined in config.py and with policy (where [s3_bucket] is the bucket name defined in config.py):

  {
      "Version": "2012-10-17",
      "Statement": [
          {
              "Effect": "Allow",
              "Action": [
                  "logs:CreateLogGroup",
                  "logs:CreateLogStream",
                  "logs:PutLogEvents"
              ],
              "Resource": "arn:aws:logs:*:*:*"
          },
          {
              "Effect": "Allow",
              "Action": [
                  "s3:GetLifecycleConfiguration"
              ],
              "Resource": [
                  "arn:aws:s3:::[s3_bucket]"
              ]
          },
          {
              "Effect": "Allow",
              "Action": [
                  "s3:PutObject",
                  "s3:GetObject",
                  "s3:DeleteObject"
              ],
              "Resource": [
                  "arn:aws:s3:::[s3_bucket]/*"
              ]
          },
          {
              "Effect": "Allow",
              "Action": [
                  "SES:SendEmail",
                  "SES:SendRawEmail"
              ],
              "Resource": [
                  "arn:aws:ses:*:*:identity/*"
              ]
          }
      ]
  }

• creates a Lambda function with name defined in config.py using the Python 2.7 runtime, with the handler and role set appropriately