High Watermark issue with monitoring Service Manager 2012 R2

Question

Hi,
Our customer has a Service Manager / Operations Manager 2012 R2 environment where SCOM will regularly generate an alarm related to High Watermark monitoring on the Service Manager data warehouse - not every day though. The alarm looks like this:
Value of 3559 is not higher than 3559
This value hasn't changed for weeks.
Having done a lot of searching we have found nothing but reference to the same recurring flow of information where people keep mentioning different values but there is no conclusion of what this might be caused by and how to solve it. The solutions
have become outdated anyway. Here is one example:
http://social.technet.microsoft.com/forums/systemcenter/en-US/889af8f3-cdf8-4d88-ae99-5dbe18d529ef/service-manager-2012-data-warehouse-high-watermark-monitor
The Service Manager Database Account in SCOM is our service account that has sysadmin in SQL. Basically, with regards to the security configuration, our customer should be covered.
To my knowledge agent-based and agentless monitoring have both been supported since SCSM 2012 SP1. Also, we have the latest management pack for Service Manager monitoring in SCOM, at the moment this should be
7.5.3079.0.
As mentioned above the value has not been changing for quite some time and nor has it in the database.
My questions are the following:
- Is there anyone that has found a solution in the meantime, especially with regards to R2?
- What is the purpose of High Watermark monitoring for the data warehouse? Why does this need to be monitored?
- We would like to understand the nature of this value better. Should it be higher each day compared to the day before?
- Could it just be overriden in SCOM?
Many thanks in advance!

Answer 1

Rather unexpectedly, the SCOM alarm disappeared overnight and when checking the workflow in the incident that was generated from the alarm, the incident got a Resolved status right before midnight at 23:59:58. We have a ConfigMgr connector that is scheduled
to run at 23:00 which we had disabled a day ago to check something else. But I don't think it had anything to do with this. As the alarm has disappeared from the SCOM console, we have also checked the registry values. Now, instead of "3559 3559 3559"
it says "3559 3559 3585". We have had the 3559 value since 18th July, and out of a sudden something must have been triggered, causing the value to increase. What was triggered and how, no idea... Not sure either what it means that only one of the
Three numbers have been updated with a new value. When talking about the DW jobs, using the SMlets we can see every day that, MPSync, DWMaintenance and the ETL jobs are working well. What we have had though for a longer time is that the cube processing
tends to have issues once a week, where processing won't finish and the cubes get hanging in a running state. We can fix that manually every time and often we don't even have to touch DWMaintenance or the ETL jobs, we don't need to go beyond manually
working with the cube processing jobs themselves. So it might not have an impact on the ETL jobs themselves. Occasionally we do find ourselves in a situation where we we need to play around in the SQL database and manually set the required value (can't
recall if it's 3, 6 or 7) cause we get an error in the DWMaintenance batch, but even though that's an occasional issue it's not constant and also something we can solve, and furthermore it's another known issue outside of the high watermark issue. So
we are a bit uncertain but can at least see that something works sometimes that sometimes increases the high watermark value :). In our case.