mqtt-load-simulator

Simulate load on MQTT broker.

When started the simulator ramps up client number during MQTT_LOAD_RAMP_UP_SECONDS seconds (defaut is 10) until it reaches MQTT_LOAD_MIN_CLIENTS (default is 2) clients. After that clients start and stop randomly to keep its number between MQTT_LOAD_MIN_CLIENTS and MQTT_LOAD_MAX_CLIENTS (default is 2) or stay constant if these 2 parameters are equal.

Once connected, each client does a random action (publish message, subscribe to the topic, unsubscribe from the topic, do nothing) each MQTT_LOAD_CLIENT_STEP_INTERVAL milliseconds (2000 by default). Probabilities are controlled by MQTT_LOAD_CLIENT_STEP_INTERVAL, MQTT_LOAD_CLIENT_STEP_INTERVAL, MQTT_LOAD_CLIENT_STEP_INTERVAL and MQTT_LOAD_CLIENT_IDLE_PROBABILITY. Publish message payload is a random text of the dictionary words. Its length is randomly picked between MQTT_LOAD_MESSAGE_MIN_SIZE (default is 10) and MQTT_LOAD_MESSAGE_MAX_SIZE (default is 600) bytes.

Prometheus metrics are exposed on port MQTT_LOAD_MONITORING_PORT (default is 1884). For complete list of configuration parameters check src/main/resources/application.conf.

Run locally as a Docker container

Pre-requisites: - Docker

Copy toolbelt/runDocker.sh, customize the parameters you want and run it.

Run on Google Cloud Platform

Pre-requisites:

• GCP account
• Terraform (0.12) or newer installed locally

1. Download account.json from GCP Account into toolbelt/terraform_gcp directory. You will have to create a service account on GCP first if you don't have one. Choose the right roles and enable google API. If something is missing terraform let you know at a later steps and allow to retry after adding necessary roles. Once you have a Service Account, you can get account.json this way:
   • Open GCP Console in the web browser
   • Go to "IAM & admin / Service Accounts" section
   • Click "Actions / Create Key" for the specific service account, choose JSON key
   • Download .json file, rename it to account.json
   • Copy to terraform-gcp and terraform-gcp-ccloud directories`
2. Choose a GCP project or create a new one on your cloud console. Terraform will prompt you to specify your project name when applying. Write down the following information, you will need it in the next steps:
   • your GCP project name
   • your GCP region
   • your GCP zone
3. Copy configuration file example:

cd toolbelt/terraform_gcp
cp config-examples/config.auto.tfvars.example ./config.auto.tfvars

4. Customize config file config.auto.tfvars with the GCP project parameters, MQTT broker connection data and desired load settings
5. Run it with toolbelt/terraform_gcp/apply.sh. Upon successful start it will print the URL of the Grafana dashboard. Default credentials are "admin/admin"
6. When done - run toolbelt/terraform_gcp/destroy.sh to stop it

Run on Kubernetes

Pre-requisistes:

• Kubernetes cluster
• kubectl installed locally, configured with your k8s cluster credentials

1. Copy config file example:

cd toolbelt/k8s
cp config.sh example config.sh

2. Customize config file with MQTT broker location and desired load settings
3. Run the load simulator with toolbelt/k8s/apply.sh
4. When done - run toolbelt/k8s/undeploy.sh to stop it.