Server load for realtime stats?

[jit]

I'm pulling the queue status with the QM.realtime calls, but I'm wondering how much of a load I might be putting on my server.

Is there practical difference between having, say, five browsers watching the QM realtime monitoring page and five clients hitting the QM.realtime API every few seconds?

I'm assuming that there's a difference in performance hit between the realtime calls and some of the other more computational reporting.

[QueueMetrics]

It's slightly more expensive, but not so much, because all data structures are created and dropped (versus a real user session, where they are shared during the whole life-cycle) and everything is wrapped in an XML-RPC call.

It is surely less expensive than computing the same thing on six months of data!