The Crawler will run on a schedule typically once a week where it will start and crawl your site. It will use a combination of the sitemap and links extracted from the pages. If no sitemap is present it will start on the homepage and pull links to reach the rest of the site. It attempts to match templates created in the Schema App Highlighter as it parses each page. If a template is matched Schema Markup is produced and sent to Schema App servers where it is converted to JSON-LD. The JSON-LD is then stored on S3 that is attached to a CloudFront CDN. The Schema App connector in AEM will send requests on a schedule to the CDN and pull markup that is produced by the crawler, if the markup has changed it will be updated in the page and pushed to out the dispatcher.