Skip to main content

Home/ MS SQL Server and Analysis Services/ SSAS performance counters
Sergey Leifer

SSAS performance counters - 0 views

SQL Server Analysis Services Monitoring

started by Sergey Leifer on 02 Sep 10
  • Sergey Leifer
     
    SSAS performance counters


    * Direct hit ratio - Ratio of cache direct hits to cache lookups, for the period between obtaining counter values. Higher is better.
    * Evictions / sec - Rate of evictions from the cache. This is per partition per cube per database. Typically due to background cleaner. Could indicate memory issue is the cache is being excessively cleaned.
    * Misses / sec - Rate of cache misses. High ratio indicates that queries are having to go to disk
    * Current Connections - Current number of client connections established.
    * Current user sessions - Current number of user sessions established.
    * Total calculation covers - Total number of of calculation covering subcubes in MDX query plan. High value means that queries are being resolved cell by cell instead of block operations.
    * Total cells calculated - Total number of cell properties calculated. High value means queries being resolved cell by cell.
    * Total recomputes - Total number of cells recomputed due to error
    * Memory Usage KB - Memory usage of the server process. Same as perfmon counter for ProcessPrivateBytes.
    * Dimension queries / sec - high number means heavy trips to the storage engine
    * Measure group queries / sec - high number means heavy trips to the storage engine
    * Available Mbytes - Available MBytes is the amount of physical memory available to processes running on the computer
    * Page Faults / sec - high page faults / sec could indicate inadequate memory. Counter has both hard and soft page faults. Soft page faults are okay. Hard page faults require disk access and can hinder performance.
    * % Processor Time - > 80 % constant value indicates inadequate processor power
    * Avg. Disk Queue Length - disk queues occur when the disk is busy reading or writing information. High disk queue lengths indicate inadequate disk performance. Rule of thumb is Avg. Disk Queue Length < 3 for an extended period.

To Top

Start a New Topic » « Back to the MS SQL Server and Analysis Services group