eG Monitoring
 

Measures reported by SqlAzuMemGrTest

If the memory resource allocation is not done properly on the target Microsoft SQL Azure database server, query execution may get delayed due to the insufficient memory resources. The users may also experience delayed response to their queries and also the performance of the Microsoft SQL Azure database server may degrade drastically. To avoid such unpleasant situation, administrators should continuously monitor the memory utilization to accurately find out whether the memory issues are responsible for slow down in the query execution. This can be achieved using the SqlAzuMemGrTest test.

By closely monitoring the memory resources of the target Microsoft SQL Azure database server, this test reports the maximum amount of memory consumed by the query executing on the server and the number of queries that were waiting for memory. This test also reveals the query cost based on how much memory the query is about to utilize and the maximum time for which the query has to wait for memory. Using this test, administrators can easily pinpoint which query is waiting too long for execution.

Outputs of the test : One set of results for the target Microsoft SQL Azure database server being monitored.

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
Max_mem_consumed_by_quer Indicates the maximum amount of memory consumed by a query executing on this database. KB  
Query_wait_in_mem_queue Indicates the number of queries waiting for memory in this database. Number  
Max_query_cost_by_memory Indicates the maximum query cost based on how much memory the query will use on this database. Seconds The cost of an execution plan is just an estimate based on how many CPU and I/O resources that the query optimizer estimates that query will use.

A high value for this measure is a cause of concern.
Max_wait_duration_memory Indicates the maximum time for which the query had to wait for memory in this database. Seconds A high value for this measure is a cause of concern as this may sometimes indicate that the database is inaccessible.