Zabbix 3.2 VMware LLD Templates discover shared datastores multiple Times

[Starko]

Hello,
we currently try to implement the VMware Monitoring features of Zabbix. For the first three clusters I added the vCenter servers for testing and linked it with the "Template Virt VMware" and started the Pollers.

Everything works (except for Log Monitoring, but that's another topic. )

My Problem is, that the most important value would be the free space on the datastores. Since we work with vSphere Clusters containing of 2+ machines, every vSphere Host "sees" each datastore. A classic "shared storage" setups for High availability in vSphere Cluster Setups.

Now the "Template Virt VMware" creates each Hypervisor and links it with the "Template Virt VMware Hypervisor" Template. That Template has a LLD Rule for "Datastore discovery" which creates Item prototypes like "Free space on datastore {#DATASTORE} (percentage)"

That's the point where it gets interesting. Due to the fact, that each host sees each LUN aka shared storage, I have datastore Item multiplied by the amount of vsphere Hosts. See attached screenshot for clarification -> https://i.imgur.com/jCO2zeX.png

The Vmware Templates contain no Trigger Prototypes. If I would now create Trigger prototypes like: "Free disk space is less than 5% on datastore v3700_p01_00" that trigger would start X times, for each vSphere Host that is part of the cluster.

That really annoys me. I asked Google and the Zabbix forums, but no one else seems to have that problem, which really confuses me. May I do something wrong?

Anyway, I would be very interested in feedback. Thank you.

[Andreas Bollhalder]

I'm facing the same problem and it's due to the item requiring the UUID from a hypervisor. It should be possible to tell it to use the vCenter.

This is one reason, why I use vPoller.

Andreas

[Kasha]

I'm adding nnew Tag
Datastore with {#DATASTORE} as value into the trigger Low space on Vmware Datastore {#DATASTORE} within Template Virt VMware Hypervisor

Expression for trigger is
{Template Virt VMware Hypervisor:vmware.hv.datastore.size[{$URL},{HOST.HOST},{#DATASTORE},free].min(10m)}<{$HV_LOWSPACE_LIMIT:"{#DATASTORE}"}


Then create new Event Correlation
Old event tag Datastore = new event tag Datastore with Remove Action (Close new event).

After that only one alarm per datastore is triggered.
Also you may use custom limits for custom datastores.

[Starko]

Thank you very much, that sounds like a great solution!
Could you explain this in more detail or give me screenshot?

Then create new Event Correlation
Old event tag Datastore = new event tag Datastore with Remove Action (Close new event).

EDIT:
Image Trigger prototype:

https://i.imgur.com/yasyYWW.png

Image Event correlation:

https://i.imgur.com/vOml8vW.png

[Kasha]

Trigger https://imgur.com/a/3f76y
Correlation Rule https://imgur.com/dz4AAbt
Correlation Operation https://imgur.com/WEMac9E

Macro per Datastore https://imgur.com/UmbnirZ

[Starko]

Thank you very much, my screenshots were already outdated.
I will give that a try, but I think you found me a solution to a very annoying problem!
You are my hero for today. Thank you so much!

[Starko]

Trigger https://imgur.com/a/3f76y
Correlation Rule https://imgur.com/dz4AAbt
Correlation Operation https://imgur.com/WEMac9E

Macro per Datastore https://imgur.com/UmbnirZ

Thank you again so much!
Today a new Trigger fired and everything worked out very well! Only one trigger popped up and everything seemed great!
But there is always a but =( The Event history is flooded with RESOLVED Events. The first Trigger exists forever, but new Triggers are created infinitely:

https://i.imgur.com/idLC8Ff.png

Does this happen for you? I think it's about the operation: Close new event
But I have no idea how to solve it. Close old event would always create new trigger, which would also be wrong.

[Kasha]

This behavior is expected. You will see only first alert per datadtore as unresolved all another closed by correlation rule.