Alarm exludes for datastore latency

[odge]

When Veeam connects a VeeamPowerNFS, often this datastore latency is higher than our required numbers.

Well, you'd say, no problem, just add this to the alarm excludes for that datastore, and then use a nice grouping expressions to ignore those datastores starting with "Veeam".

No problem, I've done that, and I dont get the alarm anymore for the datastore. The problem is I get the alarm for Host. Because the actual error is:
"Datastore Write Latency" (xx ms) is above define threshold (yy ms) for "VeeamBackup_xxxxxx"
"The host disk (vmhba) has exceeded the threshold for total latency.

Shouldn't Veeam One correlate the two, and see that my datastore is something I have used as an excluded as well. (Even the the alarm object is for my host?)

[Vitaliy S.]

Hi Mathew,

Hmm... Makes sense, but since this is host disk latency, excluding the datastore doesn't exclude it from the host perspective. It might make sense to suppress all latency alarms for all vPower datastores, I will discuss it with our R&D team once I'm back to the office.

Thanks!

[odge]

Thanks, this will be useful, (either solution), I like the idea of having the host disk latency be able to ignore a certain datastore, because sometimes this is the case as well, you care about Host disk latency, but maybe not to:
"that older NAS/SAN"
"that capacity disk pool that is underload all day, but no one cares because its serving up non realtime data"
"Veeam Power NFS"
etc etc.

Although by default exlcuding Veeam might be a good idea, or having higher limits for Veeam.

[odge]

hi, any developments on this?

going crazy on this one

[Vitaliy S.]

Hi Matthew, I have passed this information to our R&D team, in v7 we will suppress all performance alarms for vPower datastores.

[odge]

Has this been done, because it doesn't appear to be working after I upgraded to VeeamOne (unlesss it was because the alarm was there before the upgrade), but if I resolve it it keeps coming back

[Vitaliy S.]

Hi Matthew,

Yes, it should have been done, all vPower NFS datastores are excluded from usage alarms by the datastore name mask. Can you please open a support ticket to investigate this further?

Thanks!

[odge]

Hi,

Even at host level? ("Host Datastore latency" vs "Datastore Latency")

Regards

[Vitaliy S.]

Let me check that, if it hasn't been done on the host level, then we'll add this possibility in the next updates. Thanks for the heads up!

[lymatas]

Same problem exists in v7 for VM level, during sure backup jobs:

[Veeam ONE Monitor] Alarm - VM total disk latency for xxxx_5d4ace478b0c4da6a0378a6257000643 has been changed to Error (previous state: Ok)
Details: "Datastore Highest Latency" (173 ms) is above a defined threshold (75 ms)

Thank You.

[Vitaliy S.]

Hello,

Can you please tell me if you're running Veeam ONE v7 R2 or just v7?

Thank you!

[odge]

by Vitaliy S. » Mon Sep 23, 2013 1:43 pm

Let me check that, if it hasn't been done on the host level, then we'll add this possibility in the next updates. Thanks for the heads up!

Is this in R2 now?

[veremin]

Yes, as per this sticky thread. Thanks.

[Vitaliy S.]

Moreover, you can suppress any datastore performance that is connected to the host via alarm advanced settings.