We have a VSphere environment setup in Test right now and the VMWare Engineering group is pushing back on us to deploy NPM -> VMware polling because of the following:
- Verbose SNMP messages in syslog - Fixed, they were able to filter these out
- Verbose API messages in VCenter
Every time the poller logs in and out there are informational events logged to the ESX host which then roll up to the VCenter server. This is two events every 9 minutes for each ESX host. So the issue they're having is that their logs are getting full of these messages and drowning out more useful information.
Two questions:
- Is there a way for 'us' to change this on the NPM side?
- Is anyone else experiencing this issue?
One more thing:
All along the VMware team has pressed us about the way that NPM acquires this information and they have a lot of problems with the fact that we need accounts setup on every ESX host (we have hundreds). Is there someone in the VMware NPM team that can comment on whether you will ever connect directly to the VCenter server to get all the info? (i.e.; instead of polling individual hosts)
