You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It turns out that Host Process Containers are Win32 Job Objects. It's not in the responsibility of the HCS. containerd/runc binary directly invoke the Win32 Job Object API. the hcsshim library handle the host process containers internally as job container.
Not sure, if I should add a generic collector for job objects, since Job objects are used by other software components, e.g. kubelet and MSSQL. That may confused end-users, but from Windows point of view, normal containers and hostprocess containers are using distinct Windows APIs.
Problem Statement
Currently the container metrics only collects HCS container.
https://github.com/prometheus-community/windows_exporter/blob/master/internal/collector/container/container.go#L220-L223
HPC is supported since 1.26 in Kubernetes
https://kubernetes.io/docs/tasks/configure-pod-container/create-hostprocess-pod/
HPC container not implemented in HCS container on Windows
so the user won't be able to get any metrics if the container is deployed as HPC.
Proposed Solution
No response
Additional information
No response
Acceptance Criteria
No response
The text was updated successfully, but these errors were encountered: