VM Uptime Report

[mrcreosote]

Hi.

The Uptime figures from the VM Uptime report are not what I expected to see (ie not 100%!!).

Where does Veeam One get the figures from for this calculation?

Thanks.

[Vitaliy S.]

Hi James,

VM uptime report is based on the VM heartbeat performance metric. Every VM has a certain number of heartbeats for a defined period of time. If one of your VMs had less heartbeats than it was expected for the specified time period, then this VM would have <100% of uptime in our report.

Thanks!

[mrcreosote]

Hi Vitaliy.

Thanks for the prompt response.

Cheers,

James.

[Mooncake]

Hello,

Like the question has 9 years, I'm would like to asking you on this topic if the method of collection is the same ?
Because I have VM without VMware Tools (and in the graph I don't have heartbeat information) but the VM Uptime Report can give me the number of 99.05% uptime.

Is there a way to change the metric, for exemple use the metric "OS Uptime" ? (I have to give a uptime to my customer, but for now on + 2k VMs, none have a uptime of + 99.9%)

Thanks in advance.

[nikolaj]

Hi Mooncake,

No, the calculation method hasn't changed and still relies on heartbeat counters.

Is there a way to change the metric, for exemple use the metric "OS Uptime" ? (I have to give a uptime to my customer, but for now on + 2k VMs, none have a uptime of + 99.9%)

That wouldn't be representative of the real uptime metric of an actual machine because in some cases one could be running a VM with no OS installed.

Could you clarify why installing VMware Tools is not an option for you?

Thanks.

[Mooncake]

Hello,

It's not like it's not an option, it's just that we have VMs created by client under Linux how doesn't have VMware Tools, and it's hard to force them to install it.
But it seems it working even without VMware Tools.

However, I can't use this report because I realized the uptime is impact by alarms, like for exemple a alarm for backup missing, the uptime decreases.

Thanks for you answer.

[wishr]

Hello,

Could you please clarify your statement regarding alarms?

Thanks

[Mooncake]

Hello,

Sorry for the (long) delay of my answer (giving the uptime with Veeam One was complicated because we need to prouve 99.99% SLA so 4 minutes and 22 secondes per month).

After some test my theory is : I have at least 1 missing heartbeat per month, and maybe because I'm in Advanced Scalability Deployment, the interval periode of downtime is 15 minutes. Or maybe in VM Uptime the minimum interval periode of time is 15 minutes.

Due to 15 minutes portion, I was thinking it was a consequence of alarms trigger.

Regards.

[wishr]

Hello,

It would be almost correct to say that the alarms and reports work independently, but it's also right to say that they rely on the same datasets. The Advanced scalability mode controls how frequently we are writing the data to the database. However, it's also important to note how exactly the data is saved to the database for the uptime metric. Let's say the uptime should be formed up for a single VM for 900 seconds (which applies to the Advanced scalability mode). In this case we will collect the data for the whole period of 900 seconds and calculate the uptime during this period first of all. If your VM was down for 53 seconds, the uptime for the 15 (900 sec.) min period would be 847 seconds. This value will be written to the database. Next, we will calculate the uptime for the rest of 900 sec. chunks and sum up all of the values of each of these. After that we will calculate the uptime ratio based on the entire sum. Both alarms and reports will utilise this data.

I hope this makes things a little more clear. Thank you.

[woifgaung]

Hello!

Since I am pursuing a similar question: Am I right in assuming that vMotion also has a negative effect on availability here? And what about reboots from vCenter and/or VeeamONE Server?

Regards
Wolfgang

[RomanK]

Hello Wolfgang,

I'll check the cases you mentioned with QA and update the thread.

Thanks for the interesting question.

[RomanK]

Hello Wolfgang,

Finally, I've got the test results.

vMotion does not affect the uptime.
vCenter restart changed nothing.
Host restart expectedly affects, as VMs also shut down.
Veeam ONE restart also affects as heartbeats were not collected during this time.

Thanks