This is work in progress.

The intention here is to test scalability and monitor NATS and NATS streaming during stress tests.

This assumes a clean ubuntu installation.

cd $HOME
sudo apt update
sudo apt install git
mkdir -p gopath/src/github.com/ColinSullivan1/nats-testing
git clone https://github.com/ColinSullivan1/nats-testing.git gopath/src/github.com/ColinSullivan1/nats-testing
cd gopath/src/github.com/ColinSullivan1/nats-testing/scripts
sh install_deps.sh
sudo sh setup_ub.sh

# now start monitoring
cd ../monitor/client
sh monitor.sh

Log out and back in for your shell settings (namely max fds) to take effect. On the server you wish to use for monitoring (I use my laptop), install prometheus and grafana. Alter server/prometheus.yml to use the hostname of your test machine.

sh start-prom-graf.sh

Start servers on the test machine, and from another machine, run the go applications, nats-client-sim or stan-client-sim with profiles to simulate the load you are testing. (TODO: Directions)

To monitor what is going on a grafana with prometheus and collectd is used.

The installation scripts install collectd and a prometheus exporter on the test machine.

For the monitoring machine, you'll have to install prometheus and grafana yourself.

http://<monitor host>:3000

The first time, you'll want to install and use the dashboard monitor/server/graphana-dash.json or create your own from the prometheus collectd data.

http://<test host>:9103/metrics

Ensure the exporter is reachable and generating data.

http://<monitoring host>:9090/targets

Ensure prometheus is reachable, and the target collector is UP.

• Streamline Installation (Too many steps)
• Doc client simulators