This article is primarily targeted at Czech state institutions and is based on recommendations of the Czech authority. However,...
Monitoring status of a compute clusterLukas Beran
An important process in managing compute cluster based on Microsoft HPC is also its monitoring and utilization including monitoring individual computing nodes.
To view the use of a cluster you can use cmdlet Get-HpcMetricValue. This cmdlet allows you to define different metrics that we want to monitor. I recommend to define metrics HPCJobsRunning, HPCCoresInUse, HPCTasksRunning, HPCSchedulerCores, HPCSchedulerJobs, HPCSchedulerNodes.
To view information about your computing nodes use cmdlet Get-HpcNode. Nodes can be filtered by their state, name or health.