Bug #74822 Workbench reports inconsistent statement counts (dashboard vs server status)
Submitted: 12 Nov 2014 16:06 Modified: 13 May 2016 22:54
Reporter: Jörg Brühe (OCA) Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Workbench Severity:S3 (Non-critical)
Version:6.2.3 OS:Linux (CentOS 6.6 (64 bit))
Assigned to: CPU Architecture:Any
Tags: WBBugReporter

[12 Nov 2014 16:06] Jörg Brühe 
Description:
I'm running workbench 6.2.3 on server 5.6.16.

I'm running a script that does 1,000 selects (each returning a single row) on a table specifying a non-indexed field:
   select id, sleep($SLEEP) from stm_count where inhalt like '% $CNT';
"$CNT" is the loop variable, "$SLEEP" is passed as a parameter.

I have checked the workbench report of "select per second" in both places: in the dashboard, and on the "server status" page.
In all runs, the numbers differed by a factor of 3 (three):

SLEEP   seconds   select per second    dashboard   server status
0.1     104        9.6                 10          29 / 30
0.2     204        4.9                  5          15
0.05     54       18.5                 18 / 19     56

How to repeat:
On an otherwise un-loaded server, run a loop containing a select statement and a fixed delay, and look at the "select per second" reported by the workbench.

Compare the dashboard and the "server status" page.
[12 Nov 2014 16:08] Jörg Brühe 
Test script to show bug# 74822

Attachment: stm-counters.sh (application/x-shellscript, text), 1.46 KiB.
[16 Feb 2016 11:04] Marcin Szalowicz 
Hello Jörg
Thank you for the bug report.

Verified as described.
[13 May 2016 22:54] Philip Olson 
Posted by developer:
 
Fixed as of the upcoming MySQL Workbench 6.3.7 release, and here's the changelog entry:

The Dashboard and Server Status reported different Select per second
values. The Server Status page reported incorrect (inflated by about 3x)
values.

Thank you for the bug report.