ginger - Serverless framework for Go runtime

ginger is the framework manages Serverless architecture for Go runtime.

Features

ginger manages following AWS services:

• API Gateway endpoints
• S3 storage files
• Cloudwatch schedule events and Lambda function logger
• Lambda for go1.x runtime

Requirements

• Go (we recommend latest version)
• AWS user who has above resource permissions

Installation

You can download prebuilt binary at Release. But although you need to install Go for compile Lambda function.

To see a general usage, run ginger help command.

Getting Started

Setup

Run the ginger init command at your project directory:

cd /path/to/project
ginger init
>>> some output...

If you want to use (probably almost case yes) external Go package, we suggest you should put project directory under the $GOPATH/src to enable to detect vendor tree.

For example:

export GOPATH=$HOME/go
cd $HOME/go/src/ginger-project
ginger init

ginger wants to input Lambda execution role and S3 storage name, you should input suitable value.

Create function

To create new function, run the ginger function create --name [function name] command.

ginger creates function structure under the functions/ directory, and write out to configuration of Ginger.toml.

ginger function create --name example

You can find a functions/example directory which contains main.go and Function.toml. The main.go is a Lambda function handler. The github.com/aws/aws-lambda-go/lambda is installed as default. On the other hand, the Function.toml is setting file of Lambda function e.g. memory limit, timeout, and so on.

Of course you can install additional package with go get or other verndoring tools like glide, dep, ...

Note that ginger function create creates function only on your local. To work on AWS Lambda, you need to deploy function.

Deploy function

After you modified a function, run ginger deploy function command to deploy to the AWS Lambda.

ginger deploy function (--name [destination function])

ginger compiles function automatically and archive to zip, finally send to AWS to create on destination region.

Or ginger function deploy is alias of this command, so you can also use it to deploy function.

Invoke function

Once you deployed function to AWS, you can invoke the function via AWS Lambda:

ginger function invoke --name [function name] --event [event source json]

An --event option indicates event source for input of lambda function handler. ginger gets the payload as following options and pass to the function input:

• If option doesn't exists, pass as empty payload
• If option supplied as string, pass as it is
• If option starts with @, like curl, ginger tries to load the file and pass its content

After invocation end, the result print on your terminal.

To see in details, run the help command:

ginger function help

Create Resource Endpoint

To create API endpoint, run the ginger resource create --path [endpoint path] command.

ginger creates endpoint on Ginger.toml.

ginger resource create --path /foo/bar

Note that ginger resource create creates endpoint info only on your local. To work on AWS API Gateway, you need to deploy api.

Deploy api

After you created endpoint, run ginger deploy resource command to deploy to the AWS API Gateway.

ginger deploy resource --stage [target stage] (--path [destination path])

Command creates resouce which we need, and also create root REST API if you haven't create it.

if --stage option id supplied, ginger tries to create deployment to target stage. Otherwise, only create resources.

Note that the AWS API Gateway manages endpoints as pathpart, it is part of segment, so we need to create recursively by each segment. But you don't need to care about it because ginger creates and manages sub-path automatically and save on Ginger.toml.

In detail, see AWS API Gateway documentation.

Or ginger resource deploy is alias of this command, so you can also use it to deploy resources.

Setup Lambda Integration

The AWS API Gateway supports Lambda Proxy Integration, and ginger can manage its feature.

To set up it, run ginger function mount command with function name and endpoint option:

ginger function mount

Then, ginger asks target function and endpoint by choosing list. After select both and deploy api to AWS, proxy integration creates automatically. Let's access to API Gateway URL!

Invoke endpoint

In default, the API Gateway endpoint is complicated a little. So you can invoke HTTP request through ginger resource invoke command with --stage option to determine invoke stage.

ginger resource invoke --stage [stage name]

ginger asks path input, make request URI and send HTTP request, and outputs response headers and body.

API Doc

See Command API document

Examples

Now writing...

Development

Checkout this project and build locally:

cd $GOPATH
go get github.com/ysugimoto/ginger
cd src/github.com/ysugimoto/ginger
make

On make command builds with debug flag. This flag dumps stacktrace on error and all AWS SDK requests and responses. It will help you how command processed.

We welcome your feedbacks and PRs :-)

License

MIT

Author

ysugimoto (Yoshiaki Sugimoto)