My question is relatively simple. Does the polling completion metric reflect the servers ability to complete polls strictly due to its resource availability (ie. the server has enough processor time, memory, etc. resources to do it job) or does it reflect the percentage of nodes that are able to be successfully polled (in other words - all nodes that are able to respond to my polling request did or did not reply during the last poll interval)?
The reason I ask is because I have seen some moments where my polls were 99.xx% successful but not 100%, and during that relative time frame the server didn't seem to be starving for any resources. If it is the latter - great - no problem. If it is the former, well what would I start looking at for a possible cause?