VMware Cloud Community
PaulFreedman
Enthusiast
Enthusiast
Jump to solution

VM Stress - Maintenance Schedule

Hi,

A high percentage of my VMs are alerting on high CPU causing stress, this is then having a knock on affect on capacity/time remaining and throwing my undersized machine reports out. An example is shown below;

Capturestress.PNG

The CPU demand is only peaking once a month which coincides with a scheduled AV scan, if I set a Maintenance Schedule while the scan runs, apart from stopping any alerts during that time will it also have the desired affect on my stress calculation or is there another way to go about this?

1 Solution

Accepted Solutions
Kpitt
VMware Employee
VMware Employee
Jump to solution

Paul,

Yes, that's true, however there is no option to disable peaks for Stress under the Stress badge. I believe this will also affect the Stress badge. Alternatively, you can set the Policy under Stress to use Entire Range as well (I would screenshot, but I just started the upgrade to 6.5 in my lab).

Maybe start with the Stress Entire Range change before invoking the Peak Consideration change.

-K

View solution in original post

12 Replies
erikverbruggen
Hot Shot
Hot Shot
Jump to solution

If you apply a maintenance schedule on a object, vROPS will not calculate any analytics. This should prevent the stress calcution to run.

The other option is to specify the time schedule when analytics calculations are performed but I don't know if it is possible what you want. For more information about this check this, VMware Documentation Library

Kpitt
VMware Employee
VMware Employee
Jump to solution

Hi Paul,

I would highly advise to change the policy, and un-check the box (under Virtual Machine) to not use peaks when calculating stress.

pastedImage_0.png

If you know AV scans take place and do not want to setup a schedule, this method will essentially take more of an "average" instead of basing the stress of peak values.

-K

PaulFreedman
Enthusiast
Enthusiast
Jump to solution

thanks for the reply Kpitt

Interesting setting but isn't that for capacity remaining? Un-checking would affect time and capacity remaining so rightsize calculations should be more realistic but my stress alert and badge would still be high?

Reply
0 Kudos
Kpitt
VMware Employee
VMware Employee
Jump to solution

Paul,

Yes, that's true, however there is no option to disable peaks for Stress under the Stress badge. I believe this will also affect the Stress badge. Alternatively, you can set the Policy under Stress to use Entire Range as well (I would screenshot, but I just started the upgrade to 6.5 in my lab).

Maybe start with the Stress Entire Range change before invoking the Peak Consideration change.

-K

PaulFreedman
Enthusiast
Enthusiast
Jump to solution

That's more like what i'm after, I think i'll make a few different policies with different settings to see how it affects my machines badges.

Thanks very much for the help!

PaulFreedman
Enthusiast
Enthusiast
Jump to solution

Seem to have found a nice resolution, using a combination of unselecting the Peak Consideration checkbox and also changing Stress > CPU to use the entire range rather than Peaks.

This seems to have resolved Capacity Remaining and Stress badges although I can't get Time Remaining to go green.

pastedImage_2.png

Is this just a time thing as its based on the last 12 weeks? Is there a way to reset or change the time periods? I've looked in the policies but can't see anything.

Kpitt

Reply
0 Kudos
Kpitt
VMware Employee
VMware Employee
Jump to solution

Can you attach a screenshot of the full page?

Time remaining is not something I'm too familiar with, but you could change the time buffer (within the policy, under capacity/time remaining) from 30 days to something like 1 day as a test to see if we just need to wait the duration of the buffer. Also, might be worth playing with the Data Range as well under Time Range to see if that helps at all.

The 12 weeks portion is static I believe, I don't think you can change that (or at least I haven't seen where that is set).

-K

Reply
0 Kudos
PaulFreedman
Enthusiast
Enthusiast
Jump to solution

I've previously changed the Data Range to 45 days - this one affects the time used to calculate stress and capacity remaining.

Just knocked the provisioning time buffer down to 1 day will see of it has any affect.

pastedImage_0.png

Reply
0 Kudos
Kpitt
VMware Employee
VMware Employee
Jump to solution

I believe this is going to be showing red due to the demand seen by this object.

Purely based upon the picture, the demand is going to never drop below the usuable capacity, meaning you'll be sitting at "0" for CPU forever. I would almost look at altering the policy to exclude demand (maybe instead utilize usage or allocation), or set some over commitment ratios inside vROps. If you over provision resources in vCenter or have no demand restrictions, we'll see red inside time remaining if machines are demanding more then they are actually using.

-K

Reply
0 Kudos
PaulFreedman
Enthusiast
Enthusiast
Jump to solution

Provisioning buffer doesn't seem to change anything, just took unchecked CPU from the policy and getting even stranger results, may be a bug. All greater than 1 year remaining but still showing critical.

pastedImage_0.png

Reply
0 Kudos
PaulFreedman
Enthusiast
Enthusiast
Jump to solution

May have found the answer, I think its to do with forecast rather than what my VM is demanding. Can see a huge drop when I changed the policy in Stress-free Demand but the forecast has only slightly dropped, hoping that this will level out after a period of time. I think I'll leave it a week and see how its getting on just a pity that it appears that this can't be reset.

pastedImage_3.png

Reply
0 Kudos
Kpitt
VMware Employee
VMware Employee
Jump to solution

Sounds like a plan... The only way I know to reset it is to delete the object entirely and let it re-discover :smileysilly: but that's probably not the route you want to take, as that will remove all historical data for the object.

-K

Reply
0 Kudos