Warning
You're browsing the documentation for an old version of Webiny. Consider upgrading your project to Webiny 5.40.x.
What you’ll learn
  • what does the API project application represent
  • what are the necessary cloud infrastructure resources on which the API project application relies on

The API Project Application
anchor

The API project application represents your project’s GraphQL API, which is utilized by the Admin and Website applications.

But do note that the API application is more than just a simple GraphQL interface. The default Webiny applications not only define and extend the application’s GraphQL schema, but also introduce additional processes and cloud infrastructure resources in order to achieve their goal. For example, in order to serve files and optimize images, the File Manager application brings a couple of dedicated AWS Lambda functions.

Of course, the application can additionally grow in terms of application code and cloud infrastructure on your behalf, if need be.

Diagram
anchor

For brevity, the diagram doesn’t include network-level cloud infrastructure resources, like region, VPC, availability zones, and so on. Check out the Deployment Modes section if you’re interested in that aspect of the deployed cloud infrastructure.

Note that the stateful resources like Amazon S3, Amazon Cognito, Amazon DynamoDB and Amazon ElasticSearch are deployed as part of the Core project application. These are still included in the diagram, just so it’s more clear to the reader.

Webiny Cloud Infrastructure - API - OverviewWebiny Cloud Infrastructure - API - Overview
(click to enlarge)

Description
anchor

The diagram gives an overview of the complete cloud infrastructure that is deployed by different Webiny applications, which together form the API project application.

  1. Security
  2. I18N
  3. File Manager
  4. Page Builder (with Prerendering Service)
  5. Form Builder
  6. Headless CMS

In the following text, we briefly cover how each of the different Webiny applications are utilizing the shown cloud infrastructure.

1. Security
anchor

Although the Security application does work with multiple identity providers, by default, it works with Amazon Cognito, hence its icon in the lower section of the diagram e. The identity checks are performed from the GraphQL Handler C and Headless CMS D AWS Lambda functions.With Amazon Cognito, the application also extends the GraphQL schema located on the GraphQL Handler C AWS Lambda function.For storing data, the application relies on Amazon DynamoDB F.

2. I18N
anchor

The I18N application doesn’t bring any additional cloud infrastructure resources. It just expands the GraphQL schema with its own types and resolvers, so we can say only the GraphQL Handler AWS Lambda function C gets affected by it.For storing data, the application relies on Amazon DynamoDB F.

3. File Manager
anchor

Besides extending the GraphQL schema located on the GraphQL Handler AWS Lambda function c, the File Manager application also introduces a couple of dedicated AWS Lambda functions. These perform various file-related tasks like uploading, downloading, and also image optimizations. It also introduces a single Amazon S3 bucket H, in which the files are ultimately stored.For storing data and performing search queries, the application relies on Amazon DynamoDB F and Amazon ElasticSearch Service G, respectively.

4. Page Builder (With Prerendering Service)
anchor

Besides extending the GraphQL schema, the Page Builder application also introduces a couple of AWS Lambda functions that are responsible for importing and exporting of pages.

Also, via an Amazon EventBridge (part of the Core project application, now shown on the diagram), it interacts with the prerendering service that’s deployed as part of the Website project application (more on this in the Website section.

For storing data and performing search queries, the application relies on Amazon DynamoDB F and Amazon ElasticSearch Service G, respectively.

5. Form Builder
anchor

The Form Builder application doesn’t bring any additional cloud infrastructure resources. It just extends the GraphQL schema with its own types and resolvers, so we can say only the GraphQL Handler AWS Lambda function C gets affected by it.For storing data and performing search queries, the application relies on Amazon DynamoDB F and Amazon ElasticSearch Service G, respectively.

6. Headless CMS
anchor

Besides extending the GraphQL schema located on the GraphQL Handler AWS Lambda function C , the Headless CMS application also brings an extra AWS Lambda function which is responsible for hosting the dynamic Headless CMS GraphQL API D. Note that on the diagram, it is shown “behind” the GraphQL Handler C Lambda function.For storing data and performing search queries, the application relies on Amazon DynamoDB F and Amazon ElasticSearch Service G, respectively.

Learn more about the Headless CMS API in the Headless CMS GraphQL API Overview article.

FAQ
anchor

Why Is Webiny Using AWS Lambda to Host a GraphQL Handler and Not AWS AppSync?
anchor

There are a couple of reason for that.

First, implementing a GraphQL Handler using AWS Lambda enables us and our users to make additional customizations to server’s behavior.

The default Webiny applications rely on this fact as well. For example, the I18N application gives you the ability to check what is the current locale that the user is on. The Security application gives you the ability to perform custom authentication and authorization checks inside of your resolvers. Finally, it also makes it possible for the Headless CMS application to generate different GraphQL schemas, based on the current I18N locale.

Secondly, you can test your GraphQL Handler easily with a testing library like Jestexternal link, as you would do with any other JavaScript piece of code.

Finally, this approach makes it much easier for Webiny to be hosted on not just one cloud infrastructure provider (AWS), but on many more, like Microsoft Azure or Google Cloud Platform.

Which GraphQL Implementation Are You Using?
anchor

We started with Apollo GraphQL Serverexternal link, but for Webiny version 5, we migrated to GraphQL toolsexternal link, because it’s more lightweight and both easier to implement and grasp.

Which Amazon API Gateway Is Deployed as Part of the API Application?
anchor

By default, we deploy HTTP APIsexternal link. This can be changed on your behalf, if need be.

How Is the DynamoDB Data Structured?
anchor

Webiny is following the approach called single-table designexternal link, which advocates for storing all the application data in a single DynamoDB table. So all of the Webiny applications are storing data in a single table, with distinct primary and secondary keys in order to avoid clashing of data.

Is Amazon ElasticSearch Service Deployed Into Multiple AZs?
anchor

Currently it is deployed into a single AZ, just for cost reasons. But yes, production workloads should use two or three AZs. Check out the Amazon ElasticSearch Service’s FAQ pageexternal link for more information.

Deploying the Amazon ElasticSearch Service g into multiple AZs can be achieved by adjusting the cloud infrastructure codeexternal link.