Add `requested` resources to cluster monitoring
Currently, we track the total amount of resources (CPU/Memory) and how much is being consumed. While this is helpful, we do not track how much is being "requested", or reserved in Kubernetes.
If a cluster has all of its resources requested/reserved, that can cause issues as new workloads will not be able to be started.
We should track how much CPU and Memory is being requested, as if this runs out pods will become unschedulable. We can simply add another series to the existing two charts to display this.
We should update the documentation located here: https://docs.gitlab.com/ee/user/project/clusters/#monitoring-your-kubernetes-cluster-ultimate
- Update the screenshot
- Update the text to specifically note the three metrics we collect and spell out what they are
- Usage: Resources that are being actively utilized
- Requested: Resources that have been requested
- Capacity: Total amount for the cluster
- Propose making this a line chart rather than an area chart, as usage and requested resources are independent from each other (and not cumulative).
- Propose having each line be a different color to separate the three pieces of information visually from each other more
- Propose having the capacity line be dashed to differentiate it from live data