ECS and Fargate

Run a Control Plane on AWS using Elastic Container Service (ECS) and Fargate, to set up your Private Locations and run load generators in your own AWS network

AWS Elastic Container Service (ECS) is a managed container orchestration service available on AWS. In this example:

• we use Amazon ECS to configure a service to run the Gatling Control Plane
• ECS runs our Docker containers on the AWS Fargate infrastructure
• the Control Plane configuration file is loaded from an AWS S3 bucket

This is only an example. You could, for instance, use ECS to run containers on Amazon EC2, or mount the configuration file from Amazon EFS.

S3 bucket#

In the AWS management console, from the Services menu, open S3 (or search for “S3” in the search bar). Click Create bucket.

Choose a name for the bucket, and the region where it will be stored.

Configure other options as preferred. If in doubt, keep “ACLs disabled” and “Block all public access”; we will need to allow access to the bucket with a policy.

Click Create bucket.

You can then upload your Control Plane configuration file to the bucket.

IAM role#

We need an IAM role which will allow an ECS task to:

• download the Control Plane’s configuration file stored in an S3 bucket
• spawn new load generators on EC2 when running a simulation

In the AWS management console, from the Services menu, open IAM (or search for “IAM” in the search bar). Click on Access management > Roles, and then on Create role.

Choose the “AWS service” entity type, and the “Elastic Container Service Task” use case, then click Next.

Select the policies:

• AmazonS3ReadOnlyAccess (or you can create a custom policy based on this one, to only allow access to the specific S3 bucket where you stored the Control Plane configuration)
• You can use AmazonEC2FullAccess for testing, but you will most likely want to create a more restrictive policy with only the privileges needed to spawn load generator instances on EC2:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Action": [
        "ec2:Describe*",
        "ec2:CreateTags",
        "ec2:RunInstances",
        "iam:PassRole" <1>
      ],
      "Effect": "Allow",
      "Resource": "*"
    }
  ]
}

<1> Only required when setting an instance profile on Load Generators

If you have configured an IAM instance profile in the Control plane configuration file (with the optional key iam-instance-profile), you also need to add the Actions: "iam:GetInstanceProfile", "iam:ListInstanceProfiles", and "iam:PassRole".

Then click Next.

Enter a name for the role. Add a description and tags if you want. Click on Create role.

ECS cluster#

In the AWS management console, from the Services menu, open Elastic Container Service (or search for “ECS” in the search bar). Click on Create Cluster. To run your containers on Fargate, choose a Networking Only cluster.

Then enter a cluster name (and any custom tags you may want) and click on Create.

ECS Task definition#

In the AWS management console, from the Services menu, open Elastic Container Service (or search for “ECS” in the search bar). Click on Amazon ECS > Task Definitions, then on Create new Task Definition.

Select the “Fargate” launch type compatibility and click on Next step.

Configure:

• Task definition name: a name of your choice for this task definition
• Task role: the IAM role we created earlier
• Operating system family: Linux
• Task execution role: the ecsTaskExecutionRole automatically created by AWS (or an equivalent custom role)
• Task size: the minimum values (0.5 GB of RAM and 0.25 vCPU) should be enough

Shared volume#

In the Volumes section, add a volume with a name of your choice and the “Bind mount” volume type.

Side-car container (configuration loader)#

In the Container definitions section, click on Add a container.

Configure:

• Container name: a name of your choice
• Image: amazon/aws-cli

In the Environment section:

• Essential : no
• Entry point: aws,s3,cp,s3://<bucket name>/<file name>.conf,/app/conf/control-plane.conf (replace the origin bucket name and file name with the values you used; the destination file must be /app/conf/control-plane.conf)
• Working directory: /app/conf

In the Storage and logging section, select the Mount point’s Source volume (choose the volume created earlier) and enter its Container path: /app/conf.

Click Add to add this container to the task definition.

Control plane container#

In the Container definitions section, click on Add a container.

Configure:

• Container name: a name of your choice
• Image: gatlingcorp/control-plane:latest

In the Startup dependency ordering section, add a dependency:

• Container name: the name you chose for the side-car container
• Condition: COMPLETE

In the Storage and logging section, select the Mount point’s Source volume (choose the volume created earlier) and enter its Container path: /app/conf. You can set it to read only too if desired.

Click Add to add this container to the task definition, the click Create to create the task definition.

ECS Service#

In the AWS management console, from the Services menu, open Elastic Container Service (or search for “ECS” in the search bar). Click on Amazon ECS > Clusters, then on the name of your cluster. In the Services tab, click on Create.

Configure:

• Launch type: FARGATE
• Operating system family: Linux
• Task definition: choose the task definition you created earlier
• Service name: a name of your choice for this service
• Number of tasks: 1 (to run a single instance of the control plane container)

You can modify other settings if needed for your use case, but the defaults will work. Then click Next step.

Choose the VPC, subnet(s) and security group to use for your control plane container. They must allow outbound connections to api.gatling.io on port 443, so that the control plane can retrieve tasks to execute. The other parameters are not relevant to this type of task definition. Click Next step.

You do not need to enable auto scaling for this. Click Next step.

Review your service configuration and click Create service.

The service will automatically start; you should see it in your ECS cluster’s Services tab. If you click on the service and check its Tasks tab, you should see the task’s status change as it starts. It will show the status “RUNNING” once the Control Plane container has started.

Your Control Plane is up and running!#

If you kept the default logging configuration, the control plane’s logs are sent to Amazon CloudWatch, in a log group named /ecs/<task definition name>.

After a short time, you should see your Control Plane get the “up” status in Gatling Enterprise Cloud.

You can now configure a simulation to run one one or more of this Control Plane’s pool!

Edit this page on GitHub