Running Gatling in Azure Container Instances Part 1 - Getting Gatling to run in Azure

Part 2 - Adding an API to control Gatling

I am going to preface this by saying that I know rather little about Gatling overall so this was a best effort to get this at least running. I could not find any other tutorials on how to have Gatling running at all in Azure.

I will also assume if you are reading this you have at least some knowledge of Azure and the Azure CLI.

Setting Up

If you are unaware of what Azure Container Instances are, I would suggest having a quick read through this page. Essentially they can provide a short lived container for a very cheap price.

Creating the Azure File Shares

The first thing you're going to need is an Azure Storage Account which can be either created through the portal or through a small script.

az storage account create \
    --resource-group gatling \
    --name stgatling \
    --location uksouth \
    --sku Standard_LRS

You will then need a three file shares, conf, user-files and results

az storage share create --name conf --account-name stgatling 

az storage share create --name user-files --account-name stgatling 

az storage share create --name results --account-name stgatling 

And finally you're going to want to grab the account key.

az storage account keys list --resource-group gatling --account-name stgatling --query "[0].value" --output tsv

Of course you can also grab this very simply through the portal.

Creating the Container Instance

Now you're going to want to create a YAML file to deploy the Gatling container. To do this we'll be using the container groups template but just deploying the one container. Below is an example that should work, you will just need to insert your storage key into the the volumes.

apiVersion: '2018-10-01'
location: uksouth
name: gatling-aci
properties:
  containers:
    - name: gatling
      properties:
        environmentVariables: []
        image: denvazh/gatling
        ports:
          - port: 80
        resources:
          requests:
            cpu: 1.0
            memoryInGB: 1.5
        volumeMounts:
          - mountPath: /opt/gatling/conf
            name: conf
          - mountPath: /opt/gatling/user-files
            name: user-files
          - mountPath: /opt/gatling/results
            name: results
  osType: Linux
  restartPolicy: Never
  ipAddress:
    type: Public
    ports:
      - port: 80
    dnsNameLabel: gatling-test
  volumes:
    - name: conf
      azureFile:
        sharename: conf
        storageAccountName: stgatling
        storageAccountKey: <insert-storage-key-here>
    - name: user-files
      azureFile:
        sharename: user-files
        storageAccountName: stgatling
        storageAccountKey: <insert-storage-key-here>
    - name: results
      azureFile:
        sharename: results
        storageAccountName: stgatling
        storageAccountKey: <insert-storage-key-here>

tags: {}
type: Microsoft.ContainerInstance/containerGroups

This will create and start your container group. What should happen at this point is Gatling will run but then stop very quickly as there are no simulations to run.

Uploading the configuration

So we need to go ahead and upload those to the file share user-files. There are a number of ways you can go about this depending on what OS you are using, starting off with mounting the Azure File Shares we made earlier:

• Windows - https://docs.microsoft.com/en-us/azure/storage/files/storage-how-to-use-files-windows

• Linux - https://docs.microsoft.com/en-us/azure/storage/files/storage-how-to-use-files-linux

• macOS - https://docs.microsoft.com/en-us/azure/storage/files/storage-how-to-use-files-mac

Alternatively you could upload the files through Azure Storage Explorer or through the portal.

Our folder had a simulations and resources folder directly under it.

(Ignore the bad naming)

Finally, we're going to want to take the default gatling.conf and edit two lines.

First one to edit is runDescription to be anything that isn't a blank string. Whatever your run is going to be!

Secondly, edit simulationClass to be the fully qualified class name of simulation you wish to run. For us this was apimsimulations.ApiMSimulation (we were testing out an API Management integration).

Now upload that into the conf file share.

Running the simulation

You're now all set to run the simulation.

az container start \
   --resource-group gatling
   --name gatling-aci

Container start up should be fairly quick and it will compile your simulation and run it. The easiest way to see what it is doing is just to check the logs through the portal, or you can get these from the cli.

az container logs \
   --resource-group gatling
   --name gatling-aci

At the end it should say that it has generated the report(s)

Reports generated in 2s.
Please open the following file: /opt/gatling/results/apimsimulation-20190912115211503/index.html
Global: count of failed requests is 0.0 : true
Global: mean of response time is less than 500.0 : true

And then you can retrieve those from the file share using the same method as you did to upload the files.

TODO

There are quite a few things I'd like to work out how to do, some of which I think would be fairly simple:

• Running Gatling in non-interactive mode. This should just be editing the startup command (entrypoint) for the container instance to include the simulation you wish to run. This could allow for a level of automation...
• Working out how to get the container to stop once it has finished running.
• Automating this whole process. You could use an Azure Function to create a container on demand, mounting onto the file shares. It then would just need to specify a simulation to run in non-interactive mode, wait for completion and stop / delete the container.

For a first attempt though, pretty happy with how far I managed to get it. Little cheaper and somewhat easier than spinning up a whole VM to do this!