A tool that leverages AWS to see how much client an UE4 dedicated server can handle.

Before using this tool, the following needs to be installed on both EC2 instances (via AMI) and your computer:

https://aws.amazon.com/cli/

Mac OS X

brew install awscli

Linux (Ubuntu)

sudo apt-get install python-pip

# Install AWS CLI using pip
# To upgrade an existing AWS CLI installation, use the --upgrade option:
#   sudo pip install --upgrade awscli
sudo pip install awscli

# Enable Command Completion for AWS CLI
echo 'complete -C $(which aws_completer) aws' >> ~/.bash_completion

https://stedolan.github.io/jq/

Mac OS X

brew install jq

Linux (Ubuntu)

sudo apt-get install jq

All configuration files are in conf/[env] directory.

Default configs are in conf/_default directory, they are loaded first, copy them to another directory to override the default value.

Launch server:

./launch-server.sh [env]

Launch clients:

./launch-client.sh [env] [instance-count]

Terminate server:

./terminate-server.sh [env]

Terminate clients:

./terminate-clients.sh [env]

Two custom metrics of the server are posted to AWS CloudWatch:

How many client are currently online.

How much percentage the server process utilizes the CPU (100% means occupies 1 entire core).