Ísland.is Public Web Data Flow

When a user visits https://island.is, the picture below roughly describes how the data flows through different systems and eventually gets returned back to the user as a web page.

1. We are using Cloudfront as our CDN. There we cache HTML responses, which means if a user requests a page that's found in cache, then Cloudfront will immediately answer with a cached response instead of forwarding that request to the "web".

2. Web is a Next.js server that gets it's data from the api server via GraphQL endpoints.

3. Api is a monolithic server that is essentially a gateway for Digital Iceland frontends. Developers have a choice whether they'd like to cache GraphQL endpoints and for how long that cache lasts. You can add a special ?bypass-cache=<bypass-secret> query parameter to the end of the browser url to bypass both the CDN cache and the server cache. Most data gets fetched directly from Elasticsearch but not everything is stored as a copy in Elasticsearch and is therefore fetched from Contentful directly.

4. Contentful is the CMS for the Ísland.is project and acts as the "source of truth" for our data.

5. Search-indexer is a service that syncs data from Contentful into Elasticsearch. When content changes in the CMS then a webhook calls the search-indexer and it syncs the updated data into Elasticsearch.

6. Elasticsearch is used as a search engine for the web as well as being a Contentful mirror. Content gets synced from Contentful into Elasticsearch via the search-indexer. This make the environment more resilient when Contenful experience a downage.