Autoschematic

Autoschematic is a software tool designed to support the practice of DevOps. Strictly speaking, it's a new framework for managing cloud infrastructure with a declarative infrastructure-as-code approach, much like Terraform or Pulumi. However, Autoschematic is designed from the ground up in a very different way to existing frameworks.

WhatOps?

Pardon me. When we say cloud infrastructure, we really mean any resource that serves a business need that we can provision and configure via an API. You might think of virtual networks, storage buckets, managed databases - but even more broadly, you could even consider your GMail account your personal cloud infrastructure.

When we say DevOps, we often just mean the practice of provisioning and configuring cloud infrastructure.

We say that Autoschematic is declarative, as opposed to imperative. This means that instead of working in terms of the actions you want to carry out, you work with and describe the desired state of your infrastructure.

For example, the following command is imperative; it describe an action to be taken, namely creating an AWS S3 storage bucket named "otter-pics" in the us-east-1 region.

aws s3api create-bucket \
    --bucket otter-pics \
    --region us-east-1

Likewise, you could also use the AWS console to point and click on a web-based UI and create a bucket that way. On the face of it, these are both reasonable ways to get the job done. But when working as a team, declarative methods with infrastructure-as-code become much more appealing. They allow a group of people to collaborate and agree on the desired state of infrastructure.

Let's continue with our bucket example. With Autoschematic, that bucket would look something like this:

# Filenames in Autoschematic aren't arbitrary; they serve to uniquely address and identify resources.
# Here, in the prefix "main", we have an AWS S3 bucket in the us-east-1 region named "otter-pics". 
./main/aws/s3/us-east-1/buckets/otter-pics.ron

S3Bucket(
    policy: None,
    public_access_block: PublicAccessBlock(
        block_public_acls: true,
        ignore_public_acls: true,
        block_public_policy: true,
        restrict_public_buckets: true,
    ),
    tags: {
        "project": "otter-mascot-development",
        "classification": "top-secret"
    },
)

Ok, we cheated a little. We specified quite a bit more than we did in the command-line version. But that info is important: if we had forgotten to specify our bucket correctly, our valuable otter pics may have been exposed to the world!

Now that we've written our desired state in code form, let's apply it, in order to really create the bucket.

git add main/aws/s3/us-east-1/buckets/otter-pics.ron
autoschematic apply        
╔═════════════════════════════════════════════════════════════════════════════╗
║ At main/aws/s3/us-east-1/buckets/otter-pics.ron:
║  ⟣ Create new bucket otter-pics in region us-east-1
╚═════════════════════════════════════════════════════════════════════════════╝
 ◇ Plan complete.
Type 0772 to execute all of the above actions and commit.
Hit Ctrl-c to cancel.
>0772
╔═════════════════════════════════════════════════════════════════════════════╗
║ At main/aws/s3/us-east-1/buckets/otter-pics.ron:
║  ⟖ Created S3 bucket otter-pics in region us-east-1
╚═════════════════════════════════════════════════════════════════════════════╝
 ◈ Apply succeeded! Do you wish to run git commit to track the new state? [Y/n]

Up next: Why Create Autoschematic?