While we are working hard on a next SMT version we realised that we did not publish updates that happened during the current year, let us please fix that and inform you about new features and enhancements to SMT.

Updates started this year later in June and as you can see from the list below there were a quite a lot changes. We will keep you posted about the next SMT version and we will try our best to publish the article along with actual releases.

 

Update 00.5.55

  • A new data aggregation to save some space especially for environments with a lot of execution plans
  • Updated list of wait types for collection

Update 00.5.54

  • Changes to the collection of Query Plans. Introduced two levels of granularity. The handle level for recent data is later converted to hash level.

Update: 00.5.53

  • Hotfix to align CPU reports with a collector logic updated in 00.5.51

Update: 00.5.52 

  • Moderate update determined by recent changes in 00.5.50 and 00.5.51
  • We added a report status bar to display information about displayed data, this became more demanded after we started use DWH to store older data
  • Help page was added to provide more information to the point above and how data are reported
  • Timeline graph logic was adjusted to be more precise especially when reporting data across multiple data levels
  • Enhancement to data aggregation to DWH level 0 (15-minute interval)
  • Standard minor bug fix and development to checks and reports

 Update: 00.5.51 

  • This small update fixes various post release issues we have found so far after release 00.5.50
  • Reconfigured check "CPU worker time per db" to show only difference of CPU time consumed per database instead of totals since query has been registered in plan cache.
  • Updated check "Resource governor state collection" to handle situations when it is executed on Standard Edition of SQL Service where this check is not supported.
  • Updated maintenance for SMT_Collector syntax to support newer versions of Ola scripts.
  • Fixed bug when index recommendations check was not collected if instance hosts case sensitive database.
  • Fixed bug when in rare situations same check has been executed in several threads.
  • Changed retention of deadlocks collection from 1.5 minute to 1 hour to save CPU resource in cases when SQL Instance generating lots of deadlocks
  • Minor report enhancements and fixes

Update: 00.5.50

  • DWH was introduced for most of the collected data. Three levels of aggregation above the standard collector level were added.
    •  level 0 stores 15 minute interval for 30 days
    •  level 1 stores 1 hour interval for 90 days
    •  level 2 stores data on day interval for one year period
  • Reporting returns data from multiple levels based on the selected date and interval. 
  • Added Lock type in block and deadlock report
  • Added Block duration graph in block report
  • Added charts for Resource governor reports
  • By default, SMT_Collector database runs on READ_COMMITTED_SNAPSHOT isolation level
  • Report and data collector performance 
  • Minor report enhancements and fixes

 We are looking forward to the next version and hope that so do you.

 

Kind Regards,

Michal Kovaľ

More tips and tricks

Is model db important?
by Michal Tinthofer on 15/02/2012

Few days ago I have been asked for help in process of installation a SCOM. Admins wasn’t been able to install some parts of this system because installation wizards reports some unexpected errors. When I arrived to the SQL instance I found something inte

Read more
LATCH_EX Waits
by Michal Kovaľ on 27/04/2022

Recently we had a request to optimize LATCH_EX waits on one of the production servers for our customer. Today I would like to share with you our apporach and how we handled the situation.

Read more
Collation conflict
by Michal Tinthofer on 07/03/2012

I think there is more simple and harmless way. First we need to understand the process of tempdb collation setting.

Read more