anthonycorletti/astrobase

Create kubernetes clusters quickly on GCP, AWS, or Azure.

Astrobase; Create kubernetes clusters quickly on GCP, AWS, or Azure.

---

Documentation: https://astrobase.corletti.xyz

Source Code: https://github.com/anthonycorletti/astrobase

Twitter: @astrobasecloud

---

Astrobase is best for developers who create and manage reproducible environments across cloud providers with Kubernetes.

The key features are:

• API First: Unlike most other infrastructure management tools, Astrobase is an API-First service; meaning you can write any client code you like to create your Kubernetes clusters.
• Kubernetes First: Astrobase only supports Kubernetes so you and your team can focus on streamlining the same application deployment story across any provider envrionment you might need to run your applications on.
• Easy to use: Cluster creation definitions are short and simple, and you don't have to spend hours learning a domain specific language or think about a new resource management lifecycle. Astrobase only does what cloud providers do.
• Start simple: Astrobase's simplest example takes about 5 minutes to complete.
• Scale across clouds: If you're using Astrobase, and shipping your software to customers that use different cloud providers, you can test your deployments seamlessly and take advantage of over $300,000 in cloud provider credits while doing so.

Requirements

Python 3.7+

Alternatively, you can run Astrobase as a docker container incase you arent using python.

Installation

pip install astrobasecloud

A Quick Example

The absolute minimum

Create a file gke-cluster.yaml that contains the following content.

---
cluster:
  name: astrobase-quickstart
  provider: gcp
  location: us-central1-c
  node_pools:
    - name: default
      initial_node_count: 1
      autoscaling:
        enabled: true
        min_node_count: 1
        max_node_count: 3

Create a project on Google Cloud and link a billing account to the new project.

PROJECT_ID=ab-quickstart-$(date +%s)
gcloud projects create ab-quickstart-$(date +%s)
gcloud config set project $PROJECT_ID

Deploy

Start the astrobase server in one terminal session

astrobase server

Create your first profile. A profile points your cli to a particular astrobase server.

astrobase profile create local --no-secure \
export ASTROBASE_PROFILE=local

In another session, setup your GCP project and deploy your cluster!

astrobase provider setup gcp \
--project-id $(gcloud config get-value project) \
--service-name "container.googleapis.com"

astrobase cluster gke create \
--project-id $(gcloud config get-value project) \
--file "gke-cluster.yaml"

Done!

Download your credentials and make a request to the cluster once it's in a ready state

gcloud container clusters \
get-credentials astrobase-quickstart \
--zone us-central1-c && \
kubectl get nodes

Now it's time to clean-up.

astrobase cluster gke delete \
--project-id $(gcloud config get-value project) \
--file "gke-cluster.yaml"
gcloud projects delete $PROJECT_ID

Going Multi-Cloud

Two clusters, different clouds

Let's see what it takes to deploy onto two environments using Astrobase. Let's use GCP and AWS for this example.

Create a file gke-cluster.yaml with:

---
cluster:
  name: astrobase-quickstart
  provider: gcp
  location: us-central1-c
  node_pools:
    - name: default
      initial_node_count: 1
      autoscaling:
        enabled: true
        min_node_count: 1
        max_node_count: 3

Now create a file eks-cluster.yaml with:

---
cluster:
  name: astrobase-quickstart
  provider: eks
  region: us-east-1
  nodegroups:
    - nodegroupName: default
      scalingConfig:
        desiredSize: 1
        minSize: 1
        maxSize: 3

Deploy

Start the astrobase server in one terminal session

astrobase server

In another session, setup your GCP project and deploy your cluster!

astrobase provider setup gcp \
--project-id $(gcloud config get-value project) \
--service-name "container.googleapis.com"

astrobase cluster gke create \
--project-id $(gcloud config get-value project) \
--file "gke-cluster.yaml"

Then deploy your AWS EKS cluster!

astrobase cluster eks create \
--kubernetes-control-plane-arn=$(aws iam list-roles | jq -r '.Roles[] | select(.RoleName == "AstrobaseEKSRole") | .Arn') \
--cluster-subnet-id=$(aws ec2 describe-subnets --query 'Subnets[].SubnetId[]' | jq -r '.[0]') \
--cluster-subnet-id=$(aws ec2 describe-subnets --query 'Subnets[].SubnetId[]' | jq -r '.[1]') \
--cluster-security-group-id=$(aws ec2 describe-security-groups --query 'SecurityGroups[].GroupId' | jq -r '.[0]') \
--nodegroup-noderole-mapping="default=$(aws iam list-roles | jq -r '.Roles[] | select(.RoleName == "AstrobaseEKSNodegroupRole") | .Arn')" \
--file "eks-cluster.yaml"

Deploying your EKS cluster requires a little extra setup. Checkout the AWS user guide section for more details.

Now it's time to clean-up.

astrobase cluster gke delete \
--project-id $(gcloud config get-value project) \
--file "gke-cluster.yaml"

astrobase cluster eks delete \
--kubernetes-control-plane-arn=$(aws iam list-roles | jq -r '.Roles[] | select(.RoleName == "AstrobaseEKSRole") | .Arn') \
--cluster-subnet-id=$(aws ec2 describe-subnets --query 'Subnets[].SubnetId[]' | jq -r '.[0]') \
--cluster-subnet-id=$(aws ec2 describe-subnets --query 'Subnets[].SubnetId[]' | jq -r '.[1]') \
--cluster-security-group-id=$(aws ec2 describe-security-groups --query 'SecurityGroups[].GroupId' | jq -r '.[0]') \
--nodegroup-noderole-mapping="default=$(aws iam list-roles | jq -r '.Roles[] | select(.RoleName == "AstrobaseEKSNodegroupRole") | .Arn')" \
--file "eks-cluster.yaml"

Recap

In summary, Astrobase makes it incredibly simple to create multiple kubernetes environments in different cloud providers.

You don't have to learn a new language, you can extend the api if you need, deploy Astrobase into your cloud architecture, or simply run it locally.

For a more complete example including more features and detail, continue reading the user guide.

License

This project is licensed under the Apache 2.0 License.