Source files for Docker image used to send GCE Custom metrics for 5m load average and root volume utilisation.
We want to have load average and disk utilisation on the monitoring dashboard, but the Google Ops Agent sends too much data, and Flatcar doesn't support it anyway.
This project uses Python3 in order to incorporate the
google-cloud-monitoring
pip library.
Scripts are based on Getting started with Google Cloud Monitoring by Arpana Mehta
The Dockerfile
is built using:
sudo docker build -t atsigncompany/at_swarm_load .
and then pushed to Docker Hub with:
sudo docker push atsigncompany/at_swarm_load
it can then be installed and run on Swarm nodes with:
sudo docker run -d --restart unless-stopped atsigncompany/at_swarm_load
A version has also been created for AWS VMs. It needs a service account key as a .json file to replace the placeholder.
First install dependencies:
sudo apt install python3-pip
pip3 install --no-cache-dir google-cloud-monitoring
Then copy over the files in the aws directory to their respective places on
the Ubuntu VM filesystem. sudo
will be needed to copy the systemd service
definition into place.
Finally enable and start the service:
sudo systemctl enable gcp-mon.service
sudo systemctl start gcp-mon.service
Created by @cpswan