eG Monitoring
 

Measures reported by Exc2013DCConTest

Many client processes/services on the Exchange server interact with Active Directory to obtain useful recipient and configuration information. For example, you have the Microsoft.Exchange.EdgeSyncSvc process that keeps the recipient and configuration information up to date when an Edge Server is subscribed to the same AD site as the Mailbox server. You also have the ADTopologyService that locates Active Directory domain controllers and global catalog servers and provides Active Directory topology information to other Exchange Server services.

When the communication between any of these processes/services and AD slow down, user experience with Exchange will certainly be impacted adversely. If a user then complains that Exchange is slow, administrators will have to instantly figure out which process's interactions with AD are abnormal and where is the bottleneck – in running LDAP search queries? In processing LDAP read requests? In processing LDAP write requests? The Exc2013DCConTest test provides administrators with these insights.

This test auto-discovers the critical client processes/services running on the Exchange server and reports how quickly every process services the LDAP search/read/write requests it receives. In the process, the test accurately pinpoints where the bottleneck is.

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
Ldap_read_cal Indicates the rate at which Depth 0 LDAP search calls are made by this process. Calls/Sec Depth 0 calls refer to search queries that search only the base DN.

Compare the value of this measure across processes to identify which process is imposing the maximum read request load on the AD server.

Use the detailed diagnosis of this measure to know which instances of a given process are currently communicating with the AD server, and the health of the interactions of each instance.

Ldap_read_tim Indicates the time taken by this process to execute LDAP read requests and return a response. Secs The average value for this measure should be less than 0.05 seconds. Spikes (Maximum) should not be higher than 0.1 seconds.

Compare the value of this measure across processes to identify which process is processing read requests slowly. Once the process is identified, compare the Ldap_read_tim of that process with the value of the Ldap_srch_tim and Ldap_write_tim measures to know where exactly the slowdown occurred – when processing search requests, read requests, or write requests?

Ldap_srch_cal Indicates the rate at which Depth 1 and 2 LDAP search calls were made by this process. Calls/Sec Depth 1 and 2 calls refer to search queries that search 1 and 2 levels below the base DN.

Compare the value of this measure across processes to identify which process is imposing the maximum search request load on the AD server.

Ldap_srch_tim Indicates the time taken by this process to send run an LDAP search query on AD and receive a response. Secs The average value for this measure should be less than 0.05 seconds. Spikes (Maximum) should not be higher than 0.1 seconds.

Compare the value of this measure across domain controllers to identify which process's queries are being processed very slowly by AD. Once the process is identified, compare the Ldap_srch_tim of that process with the value of the Ldap_read_tim and Ldap_write_tim measures to know where exactly the slowdown occurred – when processing search requests, read requests, or write requests?

Ldap_tim_err Indicates the number of LDAP operations made per second by this process because of an exceeded timeout. Errors/Sec  
Ldap_write_cal Indicates the rate at which this process made Add/Modify/Delete calls to AD. Calls/Sec Compare the value of this measure across processes to know which process made the maximum number of add/modify/delete calls to AD and contributed the most to its workload.
Ldap_write_tim Indicates the time taken by this process to send an add/modify/delete request to AD and receive a response. Secs A consistent increase in this value could indicate a bottleneck when adding/modifying/deleting objects in AD.

Compare the value of this measure across processes to identify which process's write requests are being processed very slowly by AD. Once the process is identified, compare the Ldap_write_tim of that process with the value of the Ldap_srch_tim and Ldap_read_tim measures to know where exactly the slowdown occurred – when processing search requests, read requests, or write requests?

Ldap_oper Indicates the number of LDAP operations made by this process per minute that took longer than the specified threshold (default threshold is 15 seconds). Operations/Minute MaxQueryDuration is an LDAP administration limit that represents the maximum time a domain controller should spend on a single search. When this limit is reached, the domain controller returns a “timeLimitExceeded” error.

By comparing the value of this measure across processes, you can identify the process that is responsible for triggering the maximum number of long-running queries.

Outstand_req Indicates the current number of pending LDAP searches for this process. Number Compare the value of this measure across processes to identify that process with the maximum number of pending search requests. The reason for this anomaly should be investigated and the source of the processing bottleneck should be cleared for optimal performance of Exchange 2013.