Support for application-specific metrics tracking

[devdattakulkarni]

Currently, KubePlus tracks the following metrics by default: CPU, memory, ingress network, egress network, and storage. We should support the ability to track application-specific metrics through ResourceMonitor CRD. The assumption here will be that the application exposes such metrics at some endpoint that KubePlus is able to scrape and surface out for ingesting into Prometheus.

[devdattakulkarni]

The changes would be to following:

• Adding a field to ResourceMonitor to define the application endpoint.
• Updating the kubectl-metrics plugin to query this endpoint for the input Kind/Instance, and surface it in the output.

To test this functionality we will need to create an application that collects some metrics and exposes them on such an endpoint.

[devdattakulkarni]

The ResourceMonitor spec looks like this:
resmonitor: apiVersion: workflows.kubeplus/v1alpha1 kind: ResourceMonitor metadata: name: hello-world-service-monitor spec: resource: kind: HelloWorldService group: platformapi.kubeplus version: v1alpha1 # This attribute indicates that Pods that are reachable through all the relationships should be used # as part of calculating the monitoring statistics. monitorRelationships: all

We can add an endpoints array to it, as follows:

resmonitor: apiVersion: workflows.kubeplus/v1alpha1 kind: ResourceMonitor metadata: name: hello-world-service-monitor spec: resource: kind: HelloWorldService group: platformapi.kubeplus version: v1alpha1 # This attribute indicates that Pods that are reachable through all the relationships should be used # as part of calculating the monitoring statistics. monitorRelationships: all app_endpoints: []

The app_endpoints will need to be accessible on the app_urls for every app instance.

kubectl metrics plugin can check if app_endpoints have been defined for the ResourceMonitor object. If there is, then it will need to get the app's url and then invoke the endpoints to get the app-specific metrics.