I have seen this a couple times… and was told about a neat way to work around the issue…. hence this post:
The current IBM Hardware MP’s 126.96.36.1991 have some unique issues…. they are running discoveries WAY too often, and also discovering properties of classes that change frequently. This fills the OpsDB and especially the Data Warehouse with discovery data, and Managed Entity Properties. To make things worse – because of the way the discoveries are written – they wont show up in the console – so we cannot easily override them to change the frequency!
Here are the MP’s in question:
What I have seen, is where they cause constant updates to discovered class instance properties… which add load to the RMS, and fill the ManagedEntityProperty table of the data warehouse.
However – when we go to the console – authoring – Object Discoveries – these discoveries are nowhere to be found!
When I load the MP into MPViewer – we can see that there are 7 discoveries in this MP. Look at the frequency. This frequency isn't itself a total big deal… as long as the discovery is lightweight and doesn't update many properties of a class with values that change frequently. Unfortunately that isn't the case.
So – what we would like to do – is to change the frequency of these discoveries, so they aren't flooding so often. It will slow down how fast we monitor new IBM systems that come online – but that is a tiny tradeoff for the impact to the management group.
Since we cannot find them in the console – here is a tip given to me by another Microsoft Engineer:
In the console – choose Tools, Search, Object Discoveries:
Type IBM in the search window, and you should find all 7 discoveries:
Highlight each discovery – and in the details pane – choose “View Knowledge”
That will launch the Discovery Property… and from there – you can create the override – and Frequency (Interval Seconds) is exposed. I am setting the 180 second and 300 second discoveries to once every 12 hours (43200 seconds) and the one main discovery, targeting Windows Computer – to 86400 seconds (once per day):
This will greatly reduce the discovery data flooding into the management group.
You can use this workaround for any MP – that has a discovery you can see in MPViewer, and cannot see in the UI.