We recently migrated our NPM web server, going from a 32bit Windows 2003 server to a 64bit Windows 2008 server. After the migration we have noticed that almost all our nodes are reporting a lot more packet loss then they were before the upgrade. Strangely, the actual response times have actually gone down after the upgrade. The database and thus all the polling settings remained the same. My understanding is that the packet loss and response time that NPM calculates is based on ICMP pings. We currently have our NPM install configured to check response time and status every 35 seconds. Does this mean it sends out one ICMP request every 35 seconds? Is anyone aware of any Windows 2008 related settings that might cause a lot more of these pings to fail? Any settings within NPM that I could change to reduce the packet loss being shown?
Matthew