The load service currently produces a load warning even when the deployment is not under CPU/Memory pressure. This should be changed based on the new node-exporter metrics for each deployment node available via the prometheus service. This also means that the service no longer needs Django. Instead, the service manager should be a simple python script that fetches the prometheus data, evaluates the load characteristics, writes a json file and then publishes that json file to nginx workers.
The load service currently produces a load warning even when the deployment is not under CPU/Memory pressure. This should be changed based on the new node-exporter metrics for each deployment node available via the prometheus service. This also means that the service no longer needs Django. Instead, the service manager should be a simple python script that fetches the prometheus data, evaluates the load characteristics, writes a json file and then publishes that json file to nginx workers.