README.md

Load testing NiFi with k6

This use case will show how to use the k6 Helm chart for a load test on NiFi. k6 will be used to send a pre-defined number of binary files in a pre-defined duration.

NiFi deployment

First, you have to install and configure NiFi and a listener processor.

a. NiFi installation

You can follow the FADI installation instructions to deploy Nifi.

b. Processor configuration

The ListenHTTP processor will be used to consume incoming data from k6.

For that NiFi processor to be reachable from a browser, we need to add a NodePort service to it. For that, navigate to the processors part of service section in values.yaml file and add this config :

processors:
    enabled: true
    ports:
      - name: listenHTTP
        port: 7001
        targetPort: 7001

Then, upgrade the NiFi release :

helm upgrade <your_release> <your_Nifi_helm_repo>

Then, you can go to the Nifi UI, login and add the ListenHTTP processor with the port number defined in the values.yaml file.

k6 deployment

First, edit the values.yaml file :

target: 
  virtualuser: 1
  iterations: 100
  duration: "10s"
  host: <your_Nifi_server_address>
  port: <your_Nifi_ListenHTTP_port>
  fileUpload: <path_to_your_file>

conf:
  schedule: "*/3 * * * *"

• virtualuser : number of virtual users simultaneously connected
• iterations : number of copies of the file sent per virtual user
• duration : period during which all files are sent
• host : the address of your Nifi server
• port : the port number of your ListenHTTP processor
• fileUpload : path to the file that will be copied and sent for the load testing
• schedule : as a cron job you need to set a schedule (every 3 min in this example)

Now, you can lauch the load test by typing this command :

helm install k6 .

To see the results, type this command :

kubectl logs <your_k6_pod>