VMware VM Snapshots Logic Module and Alert Tuning

2 hours ago, Clark_Kent said:

Fortunately, there were quit simple and handy solution. It is possible to manually use instance groups - add instances to a group then apply thresholds to the group. If you add more instances later to the group, you still have to re-apply the thresholds to the group (the group is just a way of saving time when applying the threshold, not a means to applying a policy to the group).

Yeah, instance groups are pretty limited -- it is basically creating a single manual tag on instances and as you say, they don't have any way to apply to new instances via rules or similar.  If they really were groups and supported dynamic membership based on names and properties like resource groups, then it could be helpful.  Otherwise I personally would stick with API methods since I prefer fewer surprises in my life :).

2 minutes ago, MCMCMC said:

I was able to use Filter option in snapshot datasource to filter out devices from snapshot monitoring. 

Try something like that in snapshot DS:

"##WILDVALUE##" "RegExNotMatch" "\w+_replica$"

or "##WILDALIAS##" "RegExNotMatch" "\w+_replica$" 

Just check what data is being pulled out when you hit TestScript button.

 

Sure, that works, but you are diverge from the source and updates will thwack the changes.  Or, you do this in a clone, disable the parent and then merge changes into the clone when they are published.  A lot of very manual error-prone activity, sadly.

I was able to use Filter option in snapshot datasource to filter out devices from snapshot monitoring. 

Try something like that in snapshot DS:

"##WILDVALUE##" "RegExNotMatch" "\w+_replica$"

or "##WILDALIAS##" "RegExNotMatch" "\w+_replica$" 

Just check what data is being pulled out when you hit TestScript button.

23 minutes ago, Clark_Kent said:

Hi all,

let's discuss following issue. There is a Data point AgeInHours (Age of the snapshot in hours. VMware does not recommend using a given snapshot for more than 24 to 72 hours.) However, in our case, we are completely fine to have some kind of VMs snapshots with extended aging period. Especially, we have some VMs, that are used as replica and therefore we need to store their snaphosts a little bit later. The thing is, how to adjust alerting, if we want to suppres this alert for VMs ending with "_replica"in his name and for all other VMs the default alert value could be in use. Do you have any suggestion how to manage it? Is it possible at all?

Any help will be much appreciated!

There is no rule system for thresholds -- some datasources have tuning variables (e.g., the beta SNMP_Network_Interfaces module that is in the repo), but no general rules capability for setting thresholds like this. You would need to write an API script to accomplish this by adjusting the thresholds based on the instance name or whatever other attributes you want to use (e.g., instance properties).

Hi MCMCMC,

I am not sure I could change global settings... I would just adjust only Datasource Threshold to have different alert thresholds within the one item Data Source - in our caseVMware VM Snapshots Logic Module

22 minutes ago, mnagel said:

Sure, that works, but you are diverge from the source and updates will thwack the changes.  Or, you do this in a clone, disable the parent and then merge changes into the clone when they are published.  A lot of very manual error-prone activity, sadly.

That is something I had to accept. There is no other way than API or cloning DS and loosing update capabilities.

Hi @MCMCMC and @mnagel, thank you so much for your inputs. Now, I understand, that the only way to distinguish threshold rules for particular VMs is to write API script and adjust the thresholds e.g. based on the instance name.