I'm tracking a packet loss and latency issue to three nodes that are set to Status Only polling. After reading through the various Thwack threads, it seems like polling is used primarily to determine availability statistics by sending a single ICMP packet and only following up if there's a failure. I was hoping more for a series of 5 pings (1 per second) to measure packet loss and latency a little better. If a node is actually dropping 80% of packets but the NPM poll is one that gets through, then it looks like 0% packet loss for the time period.
Is there a NPM-centric way to do this? I've read that some people lower their polling interval down to 10 seconds. I suppose I could do that since it's only three nodes that I'm looking at. At least I'm getting 6 tests a minute in this case, right?
Any other insights or methodologi
