A functionality is defined that receives a post request and stores the request.
However, this functionality is implemented in four different ways to show the modular changes in the infrastructure with dev and prod configurations.
You can deploy this functionality through a pipeline too.
Clone the project
git clone https://github.com/orionverso/echoappGo to the project directory
cd echoappChoose one stack in the main function
package main
import (
"castor/stack/computesave"
"github.com/aws/aws-cdk-go/awscdk/v2"
// "github.com/aws/aws-cdk-go/awscdk/v2/awssqs"
"github.com/aws/jsii-runtime-go"
)
func main() {
defer jsii.Close()
app := awscdk.NewApp(nil)
// STACKS DIRECTLY
//computesave.NewFargateWriteToSaveBlockData(app, jsii.String("ComputeSavePipeline"), &computesave.FargateWriteToSaveBlockDataProps_DEV)
// computesave.NewFargateWriteToSaveObject(app, jsii.String("ComputeSavePipeline"), &computesave.FargateWriteToSaveObjectProps_DEV)
// computesave.NewApiWriteToSaveBlockData(app, jsii.String("ComputeSavePipeline"), &computesave.ApiWriteToSaveBlockDataProps_DEV)
computesave.NewApiWriteToSaveObject(app, jsii.String("ComputeSavePipeline"), &computesave.ApiWriteToSaveObjectProps_DEV)
app.Synth(nil)
}Generate cloudformation template
./cdk-synth-cross-account.sh <dev-region> <dev-account> <prod-region> <prod-account> --allDeploy via script
./cdk-deploy-cross-account.sh <dev-region> <dev-account> <prod-region> <prod-account> --profile <aws-user-with-admin-permision> --allExpected output
Outputs:
ComputeSavePipelineStateless1C16B3AD.ComputeApiGatewayWithLambdaProxyEndpointF7889D50 = https://<api-id>.execute-api.<region>.amazonaws.com/dev/Check funcionality
curl https://<api-id>.execute-api.<region>.amazonaws.com/dev/ \
-X POST \
-d "$RANDOM"
Thank you for take a look. I am from Lambda writing to S3. See you let's just change the storage
// STACKS DIRECTLY
//computesave.NewFargateWriteToSaveBlockData(app, jsii.String("ComputeSavePipeline"), &computesave.FargateWriteToSaveBlockDataProps_DEV)
// computesave.NewFargateWriteToSaveObject(app, jsii.String("ComputeSavePipeline"), &computesave.FargateWriteToSaveObjectProps_DEV)
computesave.NewApiWriteToSaveBlockData(app, jsii.String("ComputeSavePipeline"), &computesave.ApiWriteToSaveBlockDataProps_DEV)
//computesave.NewApiWriteToSaveObject(app, jsii.String("ComputeSavePipeline"), &computesave.ApiWriteToSaveObjectProps_DEV)Compare implementation
./cdk-diff-cross-account.sh <dev-region> <dev-account> <prod-region> <prod-account> --all
[-] AWS::S3::Bucket Stateless/Save/Bucket SaveBucket5C9EDB8F destroy
[-] AWS::S3::BucketPolicy Stateless/Save/Bucket/Policy SaveBucketPolicy255FE3E4 destroy
[-] Custom::S3AutoDeleteObjects Stateless/Save/Bucket/AutoDeleteObjectsCustomResource SaveBucketAutoDeleteObjectsCustomResourceE04EFD88 destroy
[-] AWS::IAM::Role Stateless/Custom::S3AutoDeleteObjectsCustomResourceProvider/Role CustomS3AutoDeleteObjectsCustomResourceProviderRole3B1BD092 destroy
[-] AWS::Lambda::Function Stateless/Custom::S3AutoDeleteObjectsCustomResourceProvider/Handler CustomS3AutoDeleteObjectsCustomResourceProviderHandler9D90184F destroy
[+] AWS::DynamoDB::Table Stateless/Save/Table SaveTable126A1A49
[~] AWS::IAM::Policy Stateless/Compute/FunctionWithSqsDestinations/LambdaFunction/ServiceRole/DefaultPolicy ComputeFunctionWithSqsDestinationsLambdaFunctionServiceRoleDefaultPolicy71D75A6C
Only the storage will changelet's just change the writer
// STACKS DIRECTLY
//computesave.NewFargateWriteToSaveBlockData(app, jsii.String("ComputeSavePipeline"), &computesave.FargateWriteToSaveBlockDataProps_DEV)
computesave.NewFargateWriteToSaveObject(app, jsii.String("ComputeSavePipeline"), &computesave.FargateWriteToSaveObjectProps_DEV)
//computesave.NewApiWriteToSaveBlockData(app, jsii.String("ComputeSavePipeline"), &computesave.ApiWriteToSaveBlockDataProps_DEV)
//computesave.NewApiWriteToSaveObject(app, jsii.String("ComputeSavePipeline"), &computesave.ApiWriteToSaveObjectProps_DEV)Compare implementation
./cdk-diff-cross-account.sh <dev-region> <dev-account> <prod-region> <prod-account> --all
...
[-] AWS::ApiGateway::Method Stateless/Compute/ApiGatewayWithLambdaProxy/Default/{proxy+}/ANY ComputeApiGatewayWithLambdaProxyproxyANY8A71D8E5 destroy
[-] AWS::Lambda::Permission Stateless/Compute/ApiGatewayWithLambdaProxy/Default/ANY/ApiPermission.ComputeSavePipelineStatelessComputeApiGatewayWithLambdaProxy84E1CEFD.ANY.. ComputeApiGatewayWithLambdaProxyANYApiPermissionComputeSavePipelineStatelessComputeApiGatewayWithLambdaProxy84E1CEFDANYF9700948 destroy
[-] AWS::Lambda::Permission Stateless/Compute/ApiGatewayWithLambdaProxy/Default/ANY/ApiPermission.Test.ComputeSavePipelineStatelessComputeApiGatewayWithLambdaProxy84E1CEFD.ANY.. ComputeApiGatewayWithLambdaProxyANYApiPermissionTestComputeSavePipelineStatelessComputeApiGatewayWithLambdaProxy84E1CEFDANY187A068C destroy
[-] AWS::ApiGateway::Method Stateless/Compute/ApiGatewayWithLambdaProxy/Default/ANY ComputeApiGatewayWithLambdaProxyANYC16C1EF9 destroy
[+] AWS::ElasticLoadBalancingV2::LoadBalancer Stateless/Compute/ApplicationLoadBalancedFargateService/LB ComputeApplicationLoadBalancedFargateServiceLB4A1ADA1C
[+] AWS::EC2::SecurityGroup Stateless/Compute/ApplicationLoadBalancedFargateService/LB/SecurityGroup ComputeApplicationLoadBalancedFargateServiceLBSecurityGroup03E0505C
[+] AWS::EC2::SecurityGroupEgress Stateless/Compute/ApplicationLoadBalancedFargateService/LB/SecurityGroup/to ComputeSavePipelineStatelessComputeApplicationLoadBalancedFargateServiceSecurityGroup7040C965:80 ComputeApplicationLoadBalancedFargateServiceLBSecurityGrouptoComputeSavePipelineStatelessComputeApplicationLoadBalancedFargateServiceSecurityGroup7040C96580AEE328B8
...
Only the writer will change
You can imagine the next implementationThe Fargate workloads requires a docker image in AWS Ecr. As the changes are modular, the same pipeline can be updated to serve all implementation versions.
If you are interest to execute the example cross account, you must bootstrapping first.
https://docs.aws.amazon.com/cdk/v2/guide/bootstrapping.html
You can fork this repository to trigger the pipeline on push in your own repo.
func main() {
defer jsii.Close()
app := awscdk.NewApp(nil)
// PIPELINES
// pipeline.NewPipelineDeployFargateWriteToSaveBlockData(app, jsii.String("ComputeSavePipeline"), &pipeline.PipelineDeployFargateWriteToSaveBlockDataProps_DEV)
pipeline.NewPipelineDeployFargateWriteToSaveObject(app, jsii.String("ComputeSavePipeline"), &pipeline.PipelineDeployFargateWriteToSaveObjectProps_DEV)
// pipeline.NewPipelineDeployApiWriteToSaveBlockData(app, jsii.String("ComputeSavePipeline"), &pipeline.PipelineDeployApiWriteToSaveBlockDataProps_DEV)
// pipeline.NewPipelineDeployApiWriteToSaveObject(app, jsii.String("ComputeSavePipeline"), &pipeline.PipelineDeployApiWriteToSaveObjectProps_DEV)
app.Synth(nil)
}This project was a way to show myself how powerful building in the cloud by AWS cdk can be.
Imagine that you had to rebuild an infrastructure to change the encryption of a bucket. It would be unmaintainable.
However here we can create construct patterns , reuse and change them modularly.
Finally, If your main cloud provider is AWS, I think that AWS cdk is a more than considerable option to build dynamic infrastructure that constantly adapts to new requirements.
Can general purpose programming languages define infrastructure? It's just the beginning.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent