examples

kdb+ interface for Prometheus-Exporter – example

The demonstration below is described in full here. For clarity, the following is a summary.

Requirements

This demonstration requires a Docker instance capable of running a Unix-based container e.g. Docker Desktop for Mac/Linux/Windows 19 Pro with internet access.

Setup

Start a q session locally on port 8080, running exporter.q from this folder via the command

q ../q/exporter.q -p 8080

This will expose the metrics associated with this process on port 8080 for consumption by Prometheus.

Initialize a Docker environment containing a pre-configured Prometheus and Grafana setup from within the DockerCompose folder via

Windows/macOS

Initialize the Docker instance

docker-compose up

When finished running the demonstration stop the process using Ctrl-c or run

docker-compose down

Linux

Initialize the Docker instance

docker-compose -f docker-compose-linux.yml up

Run the following when the environment is to be stopped

docker-compose -f docker-compose-linux.yml down

Example resource utilization

Provided with the interface is the script kdb_user_example.q. This can be used to show an example of resources being consumed and monitored using Prometheus. The script will connect to the q session running on port 8080 as outlined above and attempt to use resources in a number of ways

q kdb_user_example.q

Accessing Prometheus and Grafana

Once the Docker instance has been initialized, Prometheus and Grafana should be running on the following ports

• Prometheus = http://localhost:9090
• Grafana = http://localhost:3000

On the Prometheus front end you can monitor specific metrics as desired. Executing up for example will allow a user to check that the exporter is 'up'. If the demo is running correctly this will be 1 for your configured kdb+ instance.

To log into Grafana on port 3000 use the following credentials

• Username = admin
• Password = pass

Once logged in, a pre-configured dashboard named kdb+ should be available from the Home dropdown. This will give an example of monitoring, which can be be completed using the interface but is by no means exhaustive.

The following is an example of a generated dashboard from the above workflow