CPU Ready Time metric off after upgrade from 6.7 to 6.8

If that is a custom dashboard, check the properties to see of the Option is set to show the metrics as Minimum/Maximum or Average.  Min/Max will provide better visibility into volatility if that is what you are most interested in.

I verified that it is set to Minimum/Maximum and I tried other settings, but it is not reflecting vCenter numbers.  I have a case open with support, but they are having to take it back to the developers.  It is not just my custom dashboards that are incorrect, VMware Explorer is flat too.  It has worked for years and this is the first upgrade that has caused an issue like this.  I have another installation for another vSphere environment and it seems to be impacted too, but it is more difficult to tell because there were never really any regular pronounced spikes like this environment.

I created a new temporary installation of 6.7 and am colleting data from the same cluster in the pictures above.  6.7 is accurately depicting the ready time.  Here are screen captures from 6.8 (first image) and 6.7 (second image).  To limit the data, I set the filter to >.5%.  These are from the same timeframe and you will notice that 6.8 is showing a much different view than 6.7.  Both dashboards were created with the same parameters.

I see the same thing with almost everything.  I also opened a case on May 29th and they have sent it to development.  Still havn't resolved it.

6.7

6.8

My case seems to have gone into limbo.  Support has said this is a bug and a defect has been logged.  The impression that I get is that it might be fixed in the next release, but that leaves room for it might not, too.  Apparently, accurate performance monitoring data from your performance monitoring tool is not important enough to justify some sort of bug fix.  I cannot believe that they are still allowing 6.8 to be installed.  We are heading towards a critical time of the year for our business and I am not able to monitor our performance accurately unless I start over with 6.7.  This case was opened on May 10th and I get the feeling that it doesn't matter to Quest/Dell.  They have dropped the ball and really disappointed on this one.

Mike

R&D and PM are actively looking at your case. We are exploring the best way to represent what we spoke last time at the same time preserving the info presented in the chart today.

I will seperately email you to give you a better insight. Apologies for the delay, but we really want to resolve this the best way possible and thats why the time it is taking...

Thanks

Kiran

PS: I am one of the PMs on this product and we spoke about this case in mid-June timeframe

Upgraded to 7.0 and I don't see any real difference.  This is a disappointment - I waited seven months for nothing!  Hopefully I am just missing something.

Hello Mike

We added the metric VM->CPUs->summedPercentReady to replicate the behavior we had in the past, which was summing all of the per CPU percent ready values. It's available for drag and drop just like Percent Ready Time, at the same topology object level.

What is the current percentReady then? Average of all lCPUs?

I can't look at it right now, but I can tell you that I am not even looking at percent ready time today.  My new discussion has a couple of CPU utilization graphs - take a look at it if you get the chance.  I will check out the new metric tomorrow.

Thanks

Yes Thomas, the average of the individual cpus %Ready. We made this change so VMs with different numbers of CPUs can be compared based on the percentReady value. As a sum, an 8 cpu VM can look 8 times worse than a single cpu VM, when that's not actually the case.

I meant for John to reply.  ;-)

Great, I wanted to make sure we are all on the same page.  Thanks