TutorialPDF version

Add the Telemetry Tutorial NiFi Flow as a function in Cloudera DataFlow

Before you can run the Apache NiFi flow in serverless mode on AWS Lambda, you must register it in Cloudera DataFlow service and download the Lambda DF function handler libraries.

  1. After logging into the Cloudera on cloud console with the credentials that you used to register for the trial, select the Cloudera DataFlow service on the Cloudera on cloud homepage.
  2. Click Catalog in the left navigation pane to display the Flow Catalog.
  3. Click Import Flow Definition.
  4. Provide a name and a description and upload the data flow you downloaded from your local NiFi instance.
  5. Click Import.
  6. Copy the CRN # for version 1 of the flow you uploaded.
    Make sure to select the version’s CRN in the orange box below. You will need it when configuring the function in AWS Lambda.

To be able to run the NiFi flow in AWS Lambda, you need the Cloudera DataFlow function handler libraries.

  1. Click Functions in the left navigation pane and download the Cloudera DataFlow Function binaries for AWS Lambda.
    AWS Lambda will use these binaries to run the NiFi flow.
  2. Upload this binary to an S3 bucket that you will later reference when creating the function in AWS Lambda.
    The S3 bucket needs to be in the same region as where the Lambda is being created/deployed.
  3. Copy the S3 URI for later use.

You need to create a Cloudera on cloud service account for the AWS Lambda function to be able to retrieve the Cloudera DataFlow function from the Catalog where you uploaded it earlier.

  1. Go to the Cloudera on cloud homepage, select Management Console > User Management.
  2. Under Actions, select Create Machine User and provide a name for the service account.
  3. Click Roles > Update Roles and select the DFFunctionMachineUser.
  4. Click Update.
  5. Select Access Keys > Generate Access Key.
  6. Save the Access Key ID and Private Key, which will be used to configure the AWS Lambda function.

We want your opinion

How can we improve this page?

What kind of feedback do you have?