๐จ August 30, 2023: Amazon Kinesis Data Analytics has been renamed to Amazon Managed Service for Apache Flink.
โ๏ธ This blueprints application is obsolete. Please refer to the new repository
Kinesis Data Analytics Blueprints are a curated collection of Apache Flink applications. Each blueprint will walk you through how to solve a practical problem related to stream processing using Apache Flink. These blueprints can be leveraged to create more complex applications to solve your business challenges in Apache Flink, and they are designed to be extensible. We will feature examples for both the DataStream and Table API where possible.
Within this repo, you will find examples of Apache Flink applications that can be run locally, on an open source Apache Flink cluster, or on Kinesis Data Analytics Flink cluster. Clone the repository to get started.
Description | Flink API | Language |
---|---|---|
Reading from KDS and writing to Amazon S3 | DataStream | Java |
Reading from MSK Serverless and writing to Amazon S3 | DataStream | Java |
Reading from MSK Serverless and writing to MSK Serverless | DataStream | Java |
Reading from MSK Serverless and writing to Amazon S3 | Table | Python |
- In the shared CDK folder, run
npm update
. - In the CDK folder of your blueprint, run
npm update
.
For example, let's say you want to deploy the MSK to S3 blueprint. Here are the steps you would follow:
Navigate to shared CDK folder (from root of this repo)
> cd cdk-infra/shared
> npm update
up to date, audited 457 packages in 12s
30 packages are looking for funding
run `npm fund` for details
found 0 vulnerabilities
Navigate to your blueprint folder (from root of this repo)
> cd apps/java-datastream/msk-serverless-to-s3-datastream-java
> npm install
...
> npm update
up to date, audited 457 packages in 12s
30 packages are looking for funding
run `npm fund` for details
found 0 vulnerabilities
Now, you're ready to deploy blueprints.
NOTE: If npm update
doesn't actually update your dependency versions, you might have to run npm check update
or ncu
and manually update the dependency versions in the package.json
files in each of the above locations.
- To get started with a blueprint, first ensure you have the necessary prerequisites installed.
- Then clone this repo using the command shown below.
git clone https://github.com/aws-samples/amazon-kinesis-data-analytics-blueprints
- Open a terminal session and navigate to the blueprint of your choice within the project structure; once there, follow the blueprint specific instructions.
- Once you have successfully begun sending data through your blueprint, you have successfully launched and tested a blueprint!
- You can now take the blueprints in this repo, copy them to your own project structure and begin to modify them for your specific needs.