skip to content

Parsing CDK output

/ 2 min read


The Cloud Development Kit (CDK) is an IaC tool that is expected to be ran on a CI/CD environment. Although it can be successfully used on its own, the infrastructure that it creates is usually used in another step to “fully complete” a deployment.

In this post, we explore a way of parsing a CDK deployment’s output for further processing.

Our goal infrastructure

Let’s start by defining our example environment with our desired infrastructure:

  1. A Github environment running Actions upon merging to master
  2. An AWS environment containing an S3 Bucket for website hosting
  3. The Github Action calling cdk deploy to apply configuration changes
  4. CDK dumping its output to an output file, containing the bucket name
  5. A separate deploy script that parses the output file, and copying objects to the S3 bucket to complete the deployment

Steps 1-3 are out of scope of this post and we’ll focus on how to implement steps 4 and 5.

Dumping CDK output to a file

CDK is commonly used in the following manner:

$ cdk synth
$ cdk deploy --require-approval never

In order for CDK to dump an output after deployment, we can use the CfnOutput construct:

// cdk-stack.ts
const bucket = new s3.Bucket(this, 'WebsiteBucket');
new CfnOutput(this, 'WebsiteBucketName', {value: bucket.bucketName});

This will append to CDK’s output (to stdout) the following information:

MyStack.WebsiteBucketName = my-bucket-name-qe0xj1532

We can parse this as-is by piping the deploy command to a parsing tool, but we can simplify it even more by defining an outputs-file:

$ export CDK_OUTFILE=/tmp/cdk-output.json
$ cdk synth
$ cdk deploy --require-approval never --outputs-file $CDK_OUTFILE

The JSON file will look something like this:

    "MyStack": {
        "WebsiteBucketName": "my-bucket-name-qe0xj1532"

Parsing the output file

The output being formatted in a JSON file affords us a lot of convenience. We will use the jq tool to help us parse this file so we can use the bucket name as our target deployment:

#!/usr/bin/env bash


DEST_BUCKET=$(jq -r '.MyStack.WebsiteBucketName' $CDK_OUTFILE)
aws s3 sync --delete "$SOURCE_DIR" s3://"$DEST_BUCKET"/


To get output from a CDK deployment into a JSON file, use the --outputs-file flag. The JSON file can then be parsed by a tool like jq for further processing.