How Live Status works

As a user I want to see a live status on the website and this should not end up in a api call to a 3rd party site, exposing the users data.

• Twitch API has a api with web callback configurable
• ApiGateway+Lambda Process event and store it in s3
• S3 is mapped to CodePipeline
• Pipeline runs and fetch last live status document from s3
• update html in other s3 bucket
• kill cdn cache for the live status document

Links

• AWS CDK ApiGateway
• AWS CDK Lambda
• AWS CDK CodePipeline
• AWS CDK S3
• AWS CDK CodePipeline S3 source action
• Twitch Handling Webhook Events
• AWS CDK Cloudfront
• AWS S3 EventBridge
• AWS CDK EventBridge Construct Library
• Stackoverflow Authorizers dont get the body
• Blogpost Signature Validation Aplify

Design Ideas

• live status can be read on html document, which has a very small cache time set in a CDN.
• change of live status triggers codePipeline and this deploy the site with the correct status - the pipeline gets triggered by the last status file creation
• Do not use dynamodb (yet) as it is overly much to manage for little features. Keeping all simple (aka built on s3) is no crime and surely ends up less complex in terms of management.
• Managing a Site without cloudfront comes with some extra exposure and the extra cost pays for itself in terms of features (e.g. metrics).

Lambdas

• TwitchAuthorizer
• ProcessStreamEvent
• StartBuild

Events

• APIGatewayProxyEvent

• S3 Status Updated event
• S3 Website Updated event