We have Windows based worker nodes in our K8S clusters running Windows containers - it would be great if KRR also worked with the metrics that are output from the prometheus windows_exporter.
This might be as simple as using 'windows_container_cpu_usage_seconds_total' in place of 'container_cpu_usage_seconds_total' and 'windows_container_memory_usage_commit_bytes' in place of 'container_memory_working_set_bytes' though there are possibly other differences that I've not considered to complicate things.
Stretch-goal - support both sets of metric names in parallel, so I get a single report output with all of the containers in the cluster regardless of what OS they're running on.
We have Windows based worker nodes in our K8S clusters running Windows containers - it would be great if KRR also worked with the metrics that are output from the prometheus windows_exporter.
This might be as simple as using 'windows_container_cpu_usage_seconds_total' in place of 'container_cpu_usage_seconds_total' and 'windows_container_memory_usage_commit_bytes' in place of 'container_memory_working_set_bytes' though there are possibly other differences that I've not considered to complicate things.
Stretch-goal - support both sets of metric names in parallel, so I get a single report output with all of the containers in the cluster regardless of what OS they're running on.
FYI - the full list of windows container metrics is available here: https://github.com/prometheus-community/windows_exporter/blob/master/docs/collector.container.md