I have been evaluating products since the end of October, as my VM infrastructure is growing and moving to ESXi. The tools that I have looked at are:
  • vFoglight
  • vKernel Capacity Planner & Modeler
  • PerfMan for VMware
  • Veeam Monitor

I’ve also looked at Op5 Monitor’s check_esx3.pl script for creating an in-house solution. This option is of course free; but, would require heavy development time.

A great review, IMO of vFoglight is at http://vmwaretips.com/wp/2008/10/30/product-review-vizioncore-vfoglight/. I agree with the author in all aspects (have a read).

The vKernel capacity planner and Modeler aren’t truely “monitoring” solutions IMO; but, predictive analysis and further growth planning, hence their names. You can still create alarms that will indicate when capacity has been or will quickly be reached. Modeler is awesome in that, you can say “let’s add 10GB of RAM” to ESX host #1″ and modeler will calculate the metrics, state, and new capacity of the complete VM infrastructure. The price is also in-line with the competitors and they have some price incentives for 2008 year-end.

Perfman for VMware is still in development and although they can provide VI statistics and Windows/Linux counters direct from the VMs via RPC/SSH; the product appears to be over priced compared directly to the other vendors. The latest 7.2.4 build has some serious bugs that make setup difficult; however, that was not too much of a show stopper. The difficult part of PerfMan and also Veeam, is the lack of a Web interface. Perfman, for my needs is still too slow on data collection, summarization, and analytics. I also found that the alerts are not real time and need to be stripped from the PerfMan log via a third-party tool of choice. This product has the overall potential and is already there in many respects.

Veeam is unique and competitive in pricing. I was pleased to see old data already gathered from vCenter brought in immediately and data was quickly gathered in nearly a real-time interval. I noticed a 2% CPU and ~600Kbps “strain” on the vCenter server after installation on a seperate monitoring server. I see that there is another mode to run Veeam in; which will lower the traffic and slow the stats gathering. Scheduled reporting does not appear to be available in the Monitor tool; I’ll be interested in a price incentive for one of their reporting tools. I’ve honestly just started looking at Veeam, as it is the last of the four I’m evaluating.

If I can provide any other detail, please let me know!