Looking for the best approach to Service Monitoring

[steeladept]

Hello all,

I am running a nice largeish Zabbix environment monitoring a mix of Windows and Linux servers. Everything is going mostly smooth, but it has recently been asked that I start more granular monitoring of services. Up till now, we have been using the default Windows Template LLD for services to get all services. I have added a few filters to filter out a few services we don't really care to monitor, and a few chatty ones that are not really problems when they go down. However, now I am being asked to add different priorities to different services and alert on them accordingly. Moreover, we would want to (ideally) still have discovery find and identify unknown services so we can verify a new service is not added without permissions.

So the approach I am looking at was to define the services as specific items in the appropriate templates, but that raises a few questions:

1) Would that stop discovery from finding it, or would we have two entries for each of these services (one from LLD and one from the defined template)?
2) Are there any other potential drawbacks to this approach?
3) Are there better approaches you would recommend?

[steeladept]

I see several people looked at this but there are no responses so far, so I wanted to add some things I did learn in testing what I could.

The answer to #1 is no, discovery does not create entries if they already exist, so there would not be any duplication of items, and hence triggers/alerts.

So that leaves questions 2 and 3, which, of course, are subjective. I can tell anyone looking to do this that it will work. But I can't think of any other approaches, so I can't say if there is a better approach or not. If anyone does have a better approach, please feel free to weigh in below.