VMware Cloud Community
daunce
Enthusiast
Enthusiast

Difference between %RDY & %LAT_C

%RDY gets a lot of publicity, and is shown in the cpu panel of esxtop by default, but %LAT_C has to be enabled. From the description in the esx4.1 resource guide, they seem very similar.

-----------

%RDY Percentage of time the resource pool, virtual machine, or world was ready to run, but was not

provided CPU resources on which to execute.

%LAT_C Percentage of time the resource pool or world was ready to run but was not scheduled to run because

of CPU resource contention.

-----------

Although I see a lot more %LAT_C than %RDY.

Can anyone explain it in more detail and offer examples?

Cheers.

Tags (2)
0 Kudos
3 Replies
firestartah
Virtuoso
Virtuoso

Hi

As far as i've always understood it, %ready is pretty straight forward as the time a process is ready to do something but there is no cpu resources available to do the processing

%lat_c is applicable for a number of things like as it says in the example a resource pool and if there are say ten machines in a resource pool that are waiting on cpu resource for varying times, then this is averaged into a percentage and that is the latency you are getting in waiting for the CPU.

But that's my understanding and I'm sure some guru will tell em i'm wrong.

Every day's a school day :smileygrin:

Gregg

If you found this or other information useful, please consider awarding points for "Correct" or "Helpful". Gregg http://thesaffageek.co.uk
0 Kudos
daunce
Enthusiast
Enthusiast

Hi Gregg.

Thanks for the reply.

I'm still not too sure of the difference.

I haven't heard much on %LAT_C, but it sounds like it's a true indication of oversubscription, whereas %RDY may indicate 1 core free but a 4vCPU VM waiting for more cores to free up.

Any insight to your %LAT_C values?

If i make it to VMworld, i might see if i can ask the experts there.

0 Kudos
rickardnobel
Champion
Champion

daunce wrote:

whereas %RDY may indicate 1 core free but a 4vCPU VM waiting for more cores to free up.

There is another counter to watch for this, called %CSTP, where a multi vCPU has to wait for physical cores to be available.

My VMware blog: www.rickardnobel.se
0 Kudos