chmod +x singularity.sh

Mahalo and Farewell

2018-12-31T00:00:00-05:00

In a previous post, I advocated that AWS CDK is a very neat way to write infrastructure as code, enabling you to create and provision AWS infrastructure deployments predictably and repeatedly.

Today I show how to spin up a PostgreSQL RDS instance using CDK in Python. The code is available here.

Install AWS CDK

Follow theses instructions.

Create a virtual environment and install dependencies:

virtualenv venv
source venv/bin/activate
pip install -r requirements.txt

Define You RDS DB

Add any constant variable in cdk.json and then define how you want your RDS instance in postgre_sql_example/postgre_sql_example_stack.py:

class PostgreSqlExampleStack(core.Stack):

    def __init__(self, scope: core.Construct, id: str, **kwargs) -> None:
        super().__init__(scope, id, **kwargs)

        # Database Instance
        instance = rds.DatabaseInstance(self,
            'examplepostgresdbinstance', 
            master_username=master_username,
            engine=rds.DatabaseInstanceEngine.POSTGRES, instance_class=ec2.InstanceType.of(ec2.InstanceClass.BURSTABLE2, ec2.InstanceSize.MICRO), 
            vpc=self.vpc,
            auto_minor_version_upgrade=auto_minor_version_upgrade,
            availability_zone=availability_zone,
            database_name=database_name,
            enable_performance_insights=enable_performance_insights,
            storage_encrypted=storage_encrypted,
            multi_az=multi_az,
            backup_retention=backup_retention,
            monitoring_interval=monitoring_interval,
         )

Create synthesized CloudFormation templates

cdk synth

You can check what changes are introduced into your current AWS resources with:

cdk diff --profile <AWS PROFILE>

Deploy to AWS

If everything looks OK, deploy with:

cdk deploy --profile <AWS PROFILE>

To check all the stacks in the app:

cdk ls

Clean up

To destroy/remove all the newly created resources, run:

cdk destroy --profile <AWS PROFILE>

Setting up a VPC with CDK in Python

2018-08-11T09:00:00-04:00

AWS CDK is a very neat way to write infrastructure as code, enabling you to create and provision AWS infrastructure deployments predictably and repeatedly.

You choose your favorite language to code what resources (stacks) you want, and CDK synthetizes them to CloudFormation and helps you to deploy them to AWS.

In this post, I show how to set up a VPC using CDK in Python. The code is available here.

Install AWS CDK

Follow these instructions.

Create a virtual environment and install dependencies:

virtualenv venv
source venv/bin/activate
pip install -r requirements.txt

Note: If you are starting from a blank project with cdk init app --language=python instead, you will we need to manually install resources, such as pip install aws_cdk.aws_ec2.

Define You VPC

Define how you want your VPC in the file vpc_example/vpc_example_stack.py:

from aws_cdk import core, aws_ec2

class VpcExampleStack(core.Stack):

    def __init__(self, scope: core.Construct, id: str, **kwargs) -> None:
        super().__init__(scope, id, **kwargs)

        vpc = aws_ec2.Vpc(self, "MiaVPCTest", cidr="10.0.0.0/16", max_azs=3)

Create synthesized CloudFormation template

cdk synth

You can check what changes this introduces into your AWS account:

cdk diff --profile <AWS PROFILE>

Deploy to AWS

Deploy to AWS:

cdk deploy --profile <AWS PROFILE>

These are the resources that will be created with this command:

To check all the stacks in the app:

cdk ls

Clean up

To destroy/remove all the newly created resources, run:

cdk destroy --profile <AWS PROFILE>

Skaffold is a Kool K8s Tool!

2018-01-09T05:00:00-05:00

I recently found out about Skaffold, a command-line tool recently released by Google, and designed to make local Kubernetes development very easy, allowing for iterative local development against a Kubernetes cluster. It's really cool, check this out:

It runs as a binary in your local machine (e.g., easy install with brew, etc.).
It uses yaml configuration to describe how your application should be built: you specify the Docker images for your service(s) and Skaffold will tag them and push them to any valid Docker image repository.

apiVersion: skaffold/v1beta9
kind: Config
build:
  artifacts:
  - image: gcr.io/k8s-skaffold/skaffold-example
deploy:
  kubectl:
    manifests:
      - k8s-*

These images are ingested into your Kubernetes manifest and used to deploy changes to the clusters for every code change.
Skaffold watches your local development directory for filesystem changes and automatically builds and deploy your application to any local or remote Kubernetes cluster.
It is integrated to kubectl and GCP.
Skaffold not only works on your laptop as a dev tool, it also lets you reuse the same skaffold.yaml file to do deployments to your clusters in your continuous deployment system.

In resume, Skaffold makes development very easy: all you need to do is run skaffold dev (local dev workflow: Minikube, etc.) or skaffold run (continuous dev workflow: Jenkins, Travis CI, etc.) to start the entire process, creating Kubernetes pod(s) in a cluster with the Docker images of your application.

In this single command, Skaffold:

builds container images (locally or remotely),
pushes container images if the target is not local cluster,
updates Kubernetes manifests with image tags,
deploys your application with kubectl apply,
streams logs from the pods,
watches for changes in the source code and Kubernetes manifests, and then repeat 1-5.

chmod +x singularity.sh

Mahalo and Farewell

Install AWS CDK

Create a virtual environment and install dependencies:

Define You RDS DB

Create synthesized CloudFormation templates

Deploy to AWS

Clean up

Setting up a VPC with CDK in Python

Install AWS CDK

Create a virtual environment and install dependencies:

Define You VPC

Create synthesized CloudFormation template

Deploy to AWS

Clean up

Skaffold is a Kool K8s Tool!

Some learning references