During regression runs its important to collect datastore readrate and host interface transfer rates. Today the information is stored temporarily in Vsphere's Monitoring as 'Historical Metric' that are available in 5 minute intervals.

Proposed Solution:

Use Telegraf collectors with vsphere plugin to pull required metrics from vsphere api and write them to influxdb bucket.  Data will be viewable using grafana dashboard template which is already exists.  Some tweaking of the template and influx query may be necessary for certain metrics, and retention policies of data collected.

The advantages of the above is that telegraf exporter for vmware exists, the dashboard exists, and the required infra of influxdb, and telgraf would run as containers in side rhev-gw.

Links

[1] https://github.com/influxdata/telegraf/blob/master/plugins/inputs/vsphere/README.md

[2] https://tcude.net/how-to-monitor-vmware-with-grafana/