Beware. The tool mentioned here is to be used at your own risk. This is provided "AS IS" with no warranties, and confers no rights.
At a customers site AEM was in use. However, the customer decided not to use it anymore. So AEM had to be disabled.
The first steps in this process are pretty straight forward. The GPO which tells the client to forward the events to server responding to the AEM requests, has to be disabled.
Then the SCOM Management Server(s) running this feature (Client Monitoring) have to be adjusted. In the SCOM Console go to the Administration pane, Management Servers. Right click the Management Server where this feature is enabled and select ‘Disable Client Monitoring’.
Now one tends to think all is well. As a matter of a fact it is. But when one opens the SCOM Console, Monitoring pane and checks the folder ‘Agentless Exeception Monitoring’, ‘Application View’ a lot of data will be present in this View.
This is because most data is still in the Data Warehouse, and it will stay there for a long long time by default. (raw data 30 days, aggregated data 400 days).
How neat would it be when this data, and ONLY this data could be groomed out much earlier.
Running this query against the OperationsManagerDW database shows these default settings (Thanks to Kevin Holman for providing this sql-query):
Query:
SELECT AggregationTypeID, BuildAggregationStoredProcedureName, GroomStoredProcedureName, MaxDataAgeDays, GroomingIntervalMinutes
FROM StandardDatasetAggregation WHERE BuildAggregationStoredProcedureName = 'AemAggregate'
Output:
Where the grooming settings for the OpsMgr database are easily to be adjusted by the SCOM Console, this is not the case for the Data Warehouse database.
So after a bit searching I found this tool made by Daniel Savage.
And after some testing (in one of mine SCOM test environments) I changed these settings to 1 day (raw data) and 10 days (daily aggregations).
Now the output of the earlier mentioned sql-query shows this information:
This is way much better. The AEM data will be gone within two weeks!
Thanks to Daniel Savage for the tool and Kevin Holman for the query.
No comments:
Post a Comment