Playing with Serverless Cloud

A couple of weeks ago the folks behind the Serverless framework released the first version of Serverless Cloud, a managed serverless hosting service that makes it even faster to bootstrap serverless applications.

I’ve been playing around with it a bit, and love the way it is set up and how many things are provided out of the box.

The cloud CLI

Starting a new project is as simple as creating a directory and invoking the cloud CLI command.

This will bootstrap the required files, with the ability to start out with an example template.

The CLI also allows you to set up instances, run tests and seed or export data in the dedicated datastore.

Instances

By default, you’ll be working in a personal instance just for you. Everyone working on the project has their own copy of the resources in the cloud, meaning they can test and debug without the possibility of interfering with another team member’s work.

Once you’re ready to make your app available to a wider audience, you can either set up a temporary preview instance, or deploy to a specific stage (staging, production), similar to how you would deploy with the self-hosted version of the Serverless framework.

Data store

Each instance has its own built in datastore, with a handy wrapper on top of DynamoDB that allows you to easily store simple key-value pairs, as well as more complicated entities.

You can retrieve these based on labels you can assign (similar to GSIs), set expiry dates (TTL), and group items in namespaces.

It also has a very neat wrapper around DynamoDB Streams, which allow you to listen to events and filter by namespaces and keys:

data.on("created:order_*:item_*", (item) => {
   // called when an item is created that has a namespace
   // starting with "order_" and a key starting with "item_"
});

Static assets

There is also out-of-the-box support for static assets. Simply create a directory called static and put anything in there you want to be available as a static asset.

Unit and integration tests