Feature overview

One of the ultimate goals of power monitoring for Red Hat OpenShift is to provide actionable information in terms of energy consumption. 

As explained in their docs, Cost management is an OpenShift Container Platform service that enables users to better understand and track costs for clouds and containers. It is based on the upstream project Koku.

Cost management application can be accessed from cost management tab in the console: https://console.redhat.com/openshift/cost-management/ .

More info: https://www.redhat.com/en/technologies/cloud-computing/openshift/cost-management

Whether when the cost is energetic or economic, power monitoring can bring a new dimension to the data analyzed by cost management, and the goal of this Feature is to provide a first MVP for the integration between cost management and power monitoring.

Goals

Requirements

Use cases

Out of scope

To be clarified

• What's the amount of data reported by each container/hour?
• Which clouds are supported for the MVP? (AWS, GCP, Azure, on premise)
• Multicluster, ACM or single clsuter?
• To which architectures do we provide support?
• What is the outcome of the MVP? Possible outcomes:
  • Show raw consumption data
    • Per pod
    • Per cluster
  • Show cost increase over time (Energy)
  • Show cost increase over time (money)
  • Forecast future consumption
  • Optimize reousrces
  • Identify patterns that should be investigated
  • Cost integration with the rest of metrics into the cost management mix
  • Create a cost model
• Who will have access to this data? All data? Or parts of it?
• Level of support for Developer preview vs Technology Preview vs GA
• How to configure/enable this feature, since power monitoring (and related prometheus) is optional

Other background info

• Queried data today can be found here: https://github.com/project-koku/koku-metrics-operator/blob/main/collector/queries.go