VM CDP SLA compliance alarm

[mcz]

Hi guys,

I'm receiving the mentioned alarm many times per day for multiple VM's. The strange thing is that it basically tells me that CDP is out of sync or not compliant with the settings, but when I check the status in the B&R console under replicas, everything seems to be normal. Also, I've specified a time window where the alarm should get supressed, but I still get the notifications.

Where to look first? BTW, the report 'VM CDP SLA compliance ' ends with an 'internal error'.

Thanks!

[wishr]

Hi Michael,

Regarding the error when attempting to generate the report, it looks like issue #2 from the list of top known issues. It would be definitely helpful to generate the report to track down the issue.

As for the alarm itself, have you checked the statistics, as described in the KB article of the alarm? I guess there are SLA drops that caused the alarm to trigger.

Regarding the suppression rules, we are not aware of any issues with this functionality, so if you suspect there is one, I would suggest opening a support case.

Thanks

[mcz]

Fedor,

Regarding the error when attempting to generate the report, it looks like issue #2 from the list of top known issues.

Correct, thanks for that!

As for the alarm itself, have you checked the statistics, as described in the KB article of the alarm? I guess there are SLA drops that caused the alarm to trigger.

there's no KB in the article

Regarding the suppression rules, we are not aware of any issues with this functionality, so if you suspect there is one, I would suggest opening a support case.

ok, will that do later.

It would be definitely helpful to generate the report to track down the issue.

I've applied the KB and not it works. But the report tells me that we're having an average SLA of 99 % for 3 vm's and the others are at 100 %. So it's looking pretty good, so no idea why the alarm gets triggered every day over and over again...

[wishr]

Hi Michael,

Thanks for replying back!

Regarding the KB article, I meant the KB article within the alarm itself in VONE which advises where to look in VBR to find CDP policy statistics details.

Are you sure you've specified a correct reporting interval capturing the time period when the alarm got triggered? Also, what are your alarm threshold values?

Additionally, the report has a drill-down for each VM which highlights issues of each particular VM - maybe there is something suspicious mentioned there? If not, I would suggest opening a support case so we could have a look at the details because if everything is configured correctly, what you've mentioned does not look expected.

Thanks

[mcz]

Hi Fedor,

can you please double-check on your side if the alarm has a KB article? There's no KB mentioned here, just a description of the alarm, the cause and the resolution...

Looks like I haven't understood every detail yet, so please let me know the following: Are properties for the alarm only set in veeam ONE or also in B&R? Because the CDP job itself has an option to set thresholds for warnings and errors and I'm not sure which setting is the one that counts...

Thanks!

[wishr]

Hi Michael,

I apologize for the confusion, I was speaking about this alarm KB:


Alarm thresholds currently are configured independently within the alarm rule settings:

In my case, the warning will trigger if the SLA of a VM drops below 95% and the failure will arrive if the SLA goes below 75%.

Thanks

[mcz]

ok, I've created a case and got some answers today. Obviously I did not understand the alarm settings properly: The time period is not the max. allowed delay for CDP... So I guess this is why the alarm got triggered, I've done some maths and in our case the error was triggered when there was a delay of 5 minutes - which is okay for us...

Thanks!

[wishr]

Yes, there is no "delay" setting in the alarm. In our case, the time period controls the data aggregation period.

By the way, what RPO you have set within the CDP policy?

[mcz]

3 minutes short-time-retention...

[wishr]

In this case, if VBR is not able to process and inject the replicated data into the replica VM within 3 minutes, the SLA decreases in VBR for this particular VM. VONE individually tracks the SLA changes for each particular VM being processed by CDP policy and if the SLA of a VM drops below the configured threshold (in percent, which could be minimum, maximum, or average value during the configured data aggregation period), it raises the alert. That's actually why the alarm has got his current name .

[mcz]

Just one comment regarding that alarm: I've installed the latest patch yesterday and today I realized that CDP isn't working anymore, but I haven't received any alarms! Looks like the alarm doesn't get triggered when the VeeamBackupCdpSvc isn't running! Once I've started this service, veeam one triggered the alarm but not before.

Can you confirm that and shouldn't that be redesigned? Thanks!

[wishr]

Hi Michael,

The alarm we discussed earlier is intended to catch SLA drops and not issues with the services, however, it's a good idea to create a dedicated alarm for the CDP Coordinator service. Thanks for the heads up!

[mcz]

Thanks Fedor!