#4308 Research & fix timermiddleware inconsistencies

v1.0.0
closed
Cory Johns
performance (126) sf-1 (616)
General
nobody
2015-08-20
2012-05-31
Dave Brondsema
No

The 'ming' timing is always more than 'mongo'. It looks like it is measuring all the mongo query time. We should change which methods we instrument so that it is just the ming time (e.g. validation) and not include the mongo times.

The 'jinja' and 'markdown' timings often total to be more than 'total' total for a time period, which can't be correct. Having the [#4307] categorization done may help determine the cause.

We should also copy the jinja compile/render split that was done in sfpy, over to Allura. However, there is an issue since the jinja render time is consistently greater than total. So that needs to be fixed too.

Related

Tickets: #4307

Discussion

  • Dave Brondsema

    Dave Brondsema - 2012-10-09
    • milestone: forge-backlog --> forge-nov-02
     

  • Dave Brondsema

    Dave Brondsema - 2012-10-11
    • labels: --> performance
    • milestone: forge-nov-02 --> forge-oct-19
     

  • Dave Brondsema

    Dave Brondsema - 2012-10-19
    • Description has changed:

    Diff:

    --- old
+++ new
@@ -1,3 +1,5 @@
 The 'ming' timing is always more than 'mongo'.  It looks like it is measuring all the mongo query time.  We should change which methods we instrument so that it is just the ming time (e.g. validation) and not include the mongo times.

 The 'jinja' and 'markdown' timings often total to be more than 'total' total for a time period, which can't be correct.  Having the [#4307] categorization done may help determine the cause.
+
+We should also copy the jinja compile/render split that was done in sfpy, over to Allura.  However, there is an issue since the jinja render time is consistently greater than total.  So that needs to be fixed too.
     

    Related

    Tickets: #4307

  • Dave Brondsema

    Dave Brondsema - 2012-10-25

    Assuming discrepancies are due to threading issues in [#4615]

     

  • Dave Brondsema

    Dave Brondsema - 2012-10-25
    • size: --> 1
     

  • Dave Brondsema

    Dave Brondsema - 2012-11-02
    • Milestone: forge-nov-02 --> forge-nov-16
     

  • Cory Johns - 2012-11-08
    • status: open --> in-progress
    • assigned_to: Cory Johns
     

  • Cory Johns - 2012-11-08
    • status: in-progress --> blocked
     

  • Cory Johns - 2012-11-08
     

  • Cory Johns - 2012-11-08
    • status: blocked --> closed
     

Log in to post a comment.