ESX 3 & VC 2..no VMKUsage graphs, no decent at-a-glance summary graphs in VC???

[mikeyboy]

Hi please help! In VC1.3 (and VMKusage which is no longer present in ESX3), we had the ability to view performance graphs for the 4 main metrics (CPU, disk, network, mem) on one screen, for multiple hosts or VMs. For example, in VC I could click on the host, and see the 4 performance graphs for that host on one screen. I could click on the farm name, and see the 4 graphs for all the hosts in that farm, on one screen. I could click on the group name, and see the 4 graphs for all the VMs in that group, on one (or a few) screens. Point being, all the information was there one screen and I could quickly check with one click the current and past performance of my entire environment in one fell swoop. Similar with VMKUsage.

With VC2 (and no VMKUsage), it seems my only option is to have to click on one VM or host at a time to get one graph (a complicated, cluttered, confusing, and non-intuitive one at that!), which I then have to alter to show the other metrics. Do I really have to do this separately for each and every host/VM? Have I missed something? We have hundreds. I'm hoping someone will correct me because if not I will find it suddenly very difficult to monitor the performance of our large environments if that's really the case. Which would be a very unfortunate and ridiculously huge backward step. Our ability to monitor current performance and identify trends and potential capacity issues was based on the VC graphs and particularly the VMKUsage graphs (for monitoring things like CPU ready time).

Thankyou

Michael

[kjb007]

The performance tab is a simple quick at-a-glance, but when you want something a little bit more, right-click a higher-level object like a datacenter or cluster, and on the right side, click hosts, and select/highlight the hosts that you want to report, and right-click and click 'Report Performance'. Then, you can select the metrics you want, and export to a spreadsheet.

This is about the best you can do here.

Also, take a look at unnoc.org, or some commercial tools that provide similar details.

-KjB

[mikeyboy]

Hi thanks for the quick response and the suggestion, however it doesn't help me much unfortunately. I've tried the 'report performance' option, it seems that it just gives me the same per-VM or per-host individual graphs as the VI GUI does. I am unable to get it to report on multiple VMs or hosts at once (if I select multiple VMs in the right-hand window, the option to report performance is greyed out). I'd have to do it for each VM or host one-by-one, which given the amount of VMs we have, is unfeasible.

[Intex]

Not sure if this is going to give you want you want as it only applies to multiple VM's and not multiple hosts and you still have to select each host to get a view of every VM - but anyway ....

1. Select a host and go to 'Performance'

2. Select 'Chart Options'

3. Select 'CPU', 'Disk', 'Memory' etc. and then click on 'Past Day' etc. Note that you can't do this for real-time counters

4. Select 'Stacked Graph (Per VM)' in the top right-hand corner

5. Select every VM you want to report on

If that doesn't give you want you want, you could check out Vizioncore vCharter. They've just released a new version and it's pretty good. That will defintiely give you what you're after, but obviously that comes at a cost (even though it's reasonably priced per CPU socket). If it's one-off stats you're after, you could simply install the evaluation to get what you need.

With performance management being core to the health of your virtualised infrastructure, you'd think / expect these tools to be built into the OS/management toolset, but then that's rarely the case and is why there's always a third party who specialises in products in that area.

[kjb007]

If that doesn't do it for you, you'll have to go with a 3rd party tool. As mentioned by intex, vCharter is pretty good, but it is not free. Check out unnoc.org as well.

-KjB

[Intex]

<Removed - spam reply message>