Gathering of concatenated performances (e.g. datastore metrics)

[dominikduennebacke]

As previously mentioned in #22 I would like to ask for support to gather concatenated performances such as datastore or cluster metrics that are only generated every 5 minutes.

Currently vsphere-graphite does gather performances from real time metrics (sampling every 20s). The question is how to handle collection that would generate data on a lower rate that realtime (datastore, cluster and other "conceptual" metrics)

I'd suggest to provide the values of concatenated metrics together with the realtime metrics to graphite. That way realtime metrics in graphite will change every 20s whereas concatenated values will change every 5 minutes.

[cblomart]

This means doing request every minuts for the last 5 minuts metrics...