Hello, Under dT Self-mon - dT Server Stats there is a measure called "Incomplete Measurements", just wondering:
1) What classifies an incomplete measurement?
2) What could be the cause (well known) for an incomplete measurement?
Thank you for your time.
Srikar
Answer by Josh L. ·
Srikar referred me to this post as I had the same question. Christian your comment helps make more sense. I would also agree with Rob that I have seen confusion around measurements such as this.
Christian, to clarify - I should only suspect a problem if this measure increases but does not ever come back down. My initial thought on this measure was that it correlated with skipped events/purepaths but it looks like this is not the case.
Thanks,
Josh
Answer by Rob V. ·
I think the confusion on this measure seems to be caused by our "quirky overuse" of the word measure when we really mean "nodes" or "events". It's an ongoing point of confusion with customers when they try to interpret the "Measures Per Second" (really Nodes per Second) dashlet on the Server Health dashboard
Answer by Christian S. ·
hi Srikar,
the "incomplete measurements" self-mon measure returns the amount of PurePath Nodes that were entered, but not yet exited and correlated (~ number of currently running & instrumented methods).
so it's clear that this number will probably always be > 0.
so a certain amount of "incomplete measurements" does not indicate an error, as long as the number stays rather constant.
however, if this number increases statically and does not fall back, then this may indicate a problematic behavior. especially, because these incomplete nodes should be discarded after some timeout in any case.
hope this helps,
Christian
JANUARY 15, 3:00 PM GMT / 10:00 AM ET