LambdaSharp "Cebes" Release (v0.3) - 2018-09-19

Cebes was a disciple of Socrates and Philolaus, and a friend of Simmias of Thebes. He is one of the speakers in the Phaedo of Plato, in which he is represented as an earnest seeker after virtue and truth, keen in argument and cautious in decision. Xenophon says he was a member of Socrates' inner circle, and a frequent visitor to the hetaera, Theodote, in Athens.[1] He is also mentioned by Plato in the Crito and Epistle XIII. (Wikipedia)

BREAKING CHANGES

The following change may impact modules you have created using previous releases.

Bootstrap

The bootstrap procedure has been updated to include a new custom resource to handle S3 bucket subscriptions (see below). The new bootstrap procedure must be run to update the LambdaSharp environment.

Default Module Filename

The default module filename was changed from Deploy.yml to Module.yml to make it more consistent with the new terminology adopted since the v0.2 release.

A work-around to avoid being impacted by this change is to specify the module filename explicitly.

lash deploy Deploy.yml

S3 Bucket Notifications as Lambda Source

This release introduces a custom resource to handle subscribing to S3 bucket notifications. In previous releases, it was only possible to subscribe to S3 notifications for S3 buckets that were created in the same LambdaSharp module. With the addition of the S3Subscriber custom resource, it now possible to subscribe to both new and existing S3 buckets. However, this change is not backwards compatible with how previous implementation handled S3 bucket subscriptions and requires the old S3 bucket to be deleted.

LambdaSharp Environment Variable

The name of the deployment tier environment variable was changed to LAMBDASHARP_TIER. The LAMBDASHARP_PROFILE was added to allow selecting a different, default AWS profile for LambdaSharp deployments.

New LambdaSharp CLI Features

Updated New Command

The LambdaSharp CLI has a new command to create a new module file. This command creates a Module.yml file in the current directory.

To invoke the new command:

lash new module MyNewModule

Similarly, the existing new function command was updated to match.

lash new function MyFunction

Additionally, the new function command now updates the Module.yml file by adding a function definition.

Artifact Output Directory

The LambdaSharp CLI now generates all artifacts in a dedicated output directory. By default, the output directory is called bin and co-located with the input Module.yml file.

The location of the output directory can be overwritten with the --output option.

lash deploy --output MyOutputFolder

Validate LambdaSharp Assembly References

The LambdaSharp CLI now checks the version of all LambdaSharp assembly references in the function project files. If an assembly version mismatch is found, an error is emitted and deployment operation is cancelled. This behavior can be suppressed with the --skip-assembly-validation option.

lash deploy --skip-assembly-validation

To validate assembly references without deploying, use the --dryrun option.

lash deploy --dryrun

New LambdaSharp Module Features

Short-Form CloudFormation Intrinsic Functions

It is now possible to use the YAML short form for CloudFormation intrinsic functions. The short form for the following functions is now supported: !And, !Base64, !Cidr, !Equals, !FindInMap, !GetAtt, !GetAZs, !If, !ImportValue, !Join, !Not, !Or, !Ref, !Select, !Split, !Sub, !Transform.

The short form functions can be used for resources, parameters, and Lambda environment variables.

Resource

Create an SNS topic that uses the AWS region and CloudFormation stack name as display name.

Parameters:
  - Name: MyResource
    Resource:
      Type: AWS::SNS::Topic
      Properties:
        DisplayName: !Sub "${AWS::Region}-${AWS::StackName}"

Parameters

Capture the CloudFormation stack name as a Lambda parameter.

Parameters:
  - Name: MyParameter
    Value: !Ref AWS::StackName

Environment Variables

Make the AWS region available as a Lambda environment variable.

Functions:
  - Name: MyFunction
    Memory: 128
    Timeout: 30
    Environment:
        MyVariable: !Ref AWS::Region

Kinesis Stream as Lambda Source

Lambda functions can now use Kinesis streams as event sources. See Kinesis Stream sample.

Parameters:
  - Name: MyKinesisStream
    Resource:
      Type: AWS::Kinesis::Stream
      Properties:
        ShardCount: 1

Functions:
  - Name: MyFunction
    Memory: 128
    Timeout: 15
    Sources:
      - Kinesis: MyKinesisStream
        BatchSize: 15

DynamoDB Stream as Lambda Source

Lambda functions can now use DynamoDB streams as event sources. See DynamoDB Stream sample.

Parameters:
  - Name: MyDynamoDBTable
    Resource:
      Type: AWS::DynamoDB::Table
      Allow: Subscribe
      Properties:
        AttributeDefinitions:
          - AttributeName: MessageId
            AttributeType: S
        KeySchema:
          - AttributeName: MessageId
            KeyType: HASH
        ProvisionedThroughput:
          ReadCapacityUnits: 1
          WriteCapacityUnits: 1

Functions:
  - Name: MyFunction
    Memory: 128
    Timeout: 15
    Sources:
      - DynamoDB: MyDynamoDBTable
        BatchSize: 15

Export Lambda Functions

It is now possible to export a Lambda function to the parameter store in a similar way as parameters. This makes it possible for other modules to reference exported Lambda functions.

Functions:
  - Name: MyFunction
    Memory: 128
    Timeout: 15
    Export: MyExportedFunction

Collection of Lambda Parameters

Parameters can now have nested collections of parameters. This is useful when a variable number of parameters is required by a Lambda function.

Parameters:
  - Name: MyCollection
    Parameters:

      - Name: First
        Value: 1st Parameter

      - Name: Second
        Value: 2nd Parameter

The Lambda function can retrieve the names of the nested parameters by using the Keys property during initialization.

public override Task InitializeAsync(LambdaConfig config) {
    var collection = config["MyCollection"];
    foreach(var key in collection.Keys) {
        _parameters[key] = collection.ReadText(key);
    }
    return Task.CompletedTask;
}

Support CloudFormation Macros (Experimental)

This release includes experimental support for creating CloudFormation Macros. See CloudFormation Macro sample.

Note that while macros can be defined by a LambdaSharp module, they cannot be invoked by a LambdaSharp module, because they require CloudFormation stack updates to use change sets, which is not yet supported.

The following function definition creates a new macro called {{Tier}}-MyMacro that can be invoked by other CloudFormation change sets.

Functions:
  - Name: Function
    Memory: 128
    Timeout: 30
    Sources:
      - Macro: MyMacro

Fixes

  • Fixed an issue where some CloudFormation properties needed to suffixed with _ to work with Humidifier library for generating CloudFormation templates correctly.

Internal Changes

  • Improved how Lambda function parameters are passed in. Instead of relying on an embedded parameters.json file, parameters are now passed in via environment variables. This means that Lambda function packages no longer need to be re-uploaded because of parameter changes.
  • Included Sid attribute for all built-in, automatic permissions being added to provide more context.
  • Switched to native AWS Lambda JsonSerializer class for serializing/deserializing data.
  • Upgraded YamlDotNet library to 5.0.1