You receive alerts from Microsoft System Center Operations Manager (SCOM) that some Microsoft Exchange Server 2013 Managed Availability health sets are in the unhealthy state.
When you check the affected health sets in Exchange PowerShell, you find that all the health sets contain the unhealthy monitors that belong to the Microsoft.Office.Datacenter.ActiveMonitoring.OverallConsecutiveSampleValueAboveThresholdMonitor type.
For example, the Total.Shadow.Queue.Length.Above.Threshold.Monitor monitor belongs to this type and also belongs to the HubTransport health set.
Note This issue occurs when monitors of the Microsoft.Office.Datacenter.ActiveMonitoring.OverallConsecutiveSampleValueAboveThresholdMonitor type are executed on a server that has non-English format settings.
The affected monitors retrieve their sample values from performance counters through the Microsoft Exchange Diagnostics service. For example, Total.Shadow.Queue.Length.Above.Threshold.Monitor reads the MSExchangeTransport Shadow Redundancy Host Info\Shadow Queue Length\_total performance counter and compares it with a threshold value of 1,500.
For this purpose, the data is temporarily stored in .csv format in the following directory:
...\Program Files\Microsoft\Exchange Server\V15\Logging\Diagnostics\PerformanceLogsToBeProcessed In these .csv files, numeric values are stored as strings. In order to be compared with a numeric threshold value, the strings have to be converted back to a numeric format. However, this conversion is not culture agnostic. Therefore, a string such as "2.69136905670166" will be converted into the double value "269136905670166."
To resolve this issue, install Cumulative Update 8 for Exchange Server 2013.
Microsoft has confirmed that this is a problem in the Microsoft products that are listed in the "Applies to" section.