Turborepo S3 Cache

Serverless Turborepo remote caching using API Gateway and S3.

Description

This project is a serverless implementation of the Turborepo custom remote cache server.

It makes use of the preflight Turborepo feature to return signed S3 URLs to the client to upload/download artifacts directly from S3.

Build and Deployment is managed by the Serverless Framework.

Getting Started

1. Clone this repo

   git clone git@github.com:gpdenny/turborepo-s3-cache.git

2. Install dependencies

   yarn install

3. Copy example environment variables

   You can do this for each environment you might wish to deploy to, for example, dev.

   cp .env.example .env.dev

4. Modify environment variables to your requirements

• SERVICE_NAME - Serverless service name, used to prefix all AWS resources
• REGION - The AWS Region to deploy the service to
• TEAMS_TABLE_NAME - Name for the DynamoDB teams table see below
• CACHE_BUCKET_NAME - Name for the artifacts S3 bucket
• CACHE_EXPIRATION_DAYS - Number of days cache artifacts are kept in S3 before being deleted

5. Run Serverless deploy

   Include the stage flag, with the environment you wish to deploy.

   yarn serverless deploy --stage dev

6. Populate the Teams DynamoDB table

   aws dynamodb put-item \
       --table-name `TEAMS_TABLE_NAME`  \
       --item \
           '{"name": {"S": "myteam"}, "tokens": {"L": [{"S": "token_1"}, {"S": "token_2"}]}}'

7. Add remote cache to your Turborepo config

   .turbo/config.json

   {
       "teamId": "team_myteam",
       "apiUrl": "https://fxbu8kd8i.execute-api.eu-central-1.amazonaws.com/dev/"
   }

   package.json

   [...]
   "lint": "turn run lint --token'\"token1\""
   [...]

   Read More

Authors

Gareth Denny

Acknowledgments

Inspiration, code snippets, etc.

• turborepo-remote-cache
• Tommy Parnell