This site is currently read-only as we are migrating to Oracle Forums for an improved community experience. You will not be able to initiate activity until January 31st, when you will be able to use this site as normal.

    Forum Stats

  • 3,890,899 Users
  • 2,269,649 Discussions
  • 7,916,821 Comments

Discussions

elapsed_time in apex_workspace_activity_log

Keegan_W
Keegan_W Member Posts: 179 Red Ribbon
edited Feb 22, 2018 8:28PM in APEX Discussions

Trying to get a better understanding of what elapsed_time is to use it in custom charts to show our highest-used APEX apps.

The description in the apex_dictionary is "Elapsed time to generate page source"

So, if you were to sum the elapsed_time for a certain app and time period, this would be the sum of all the times any page in the app was fully loaded (in seconds)?

elapsed time.JPG

Thank you!

Keegan_WPMON

Answers

  • Scott Wesley
    Scott Wesley Systems Consultant & Trainer Perth, AustraliaMember Posts: 6,272 Gold Crown
    edited Feb 20, 2018 8:39PM

    That's how I use it. (Though it's not the entire time the user may be waiting. Networking lag...)

    And I marry this with frequency, and compared medians & averages to find those that are used often, occasionally slow, frequently slow, rarely used, etc.

    Keegan_WKeegan_W
  • Keegan_W
    Keegan_W Member Posts: 179 Red Ribbon
    edited Feb 21, 2018 9:41AM

    Thanks Scott!  Do you have an example of a query where you've married it with frequency?

  • Scott Wesley
    Scott Wesley Systems Consultant & Trainer Perth, AustraliaMember Posts: 6,272 Gold Crown
    edited Feb 21, 2018 8:27PM

    Yes, I should have clarified further. I think I used an odd choice of words

    Here I'm just looking at how often the page was used during the day.

    select to_char(view_date,'hh24') dt ,count(*) c ,min(elapsed_time) min ,round(median(elapsed_time),2) median ,round(avg(elapsed_time),2) avg ,max(elapsed_time) max ,round(avg(elapsed_time)-median(elapsed_time),2) difffrom apex_workspace_activity_logwhere view_date > sysdate-2 -- last two daysand application_id = :APP_IDand page_view_type = 'Rendering'group by to_char(view_date,'hh24')order by dt desc

    pastedImage_8.png

    I find it interesting how the outliers tend the average away from the median as the count drops.

    And you could do what some of the app builder reports do and multiple them out to find weighted averages / medians

    -- top pagesselect r.application_id||':'||r.page_id page ,p.application_name||' - '||p.page_name page ,count(*) c ,count(distinct apeX_user) unique_users ,sum(elapsed_time) sum_elapsed ,count(*) * sum(elapsed_time) weighted_sum ,count(*) * median(elapsed_time) weighted_median ,max(elapsed_time) max_elapsed ,round(avg(elapsed_time),3) avg_elapsed ,median(elapsed_time) median_elapsed ,count(case when page_view_type = 'Rendering' then 1 end) render ,count(case when page_view_type = 'AJAX' then 1 end) ajaxfrom apex_workspace_activity_log r join apx_application_Pages p  on p.page_id = r.page_id  and p.application_id = r.application_idwhere  r.application_id = :MY_APPgroup by r.application_id||':'||r.page_id ,p.application_name||' - '||p.page_name order by  c desc

    pastedImage_9.png

    In this case, I might ask why page 5 has a higher median than 1 or 3 (particularly since I'm aware of what those pages do...)

    I love this data.

    PMON
  • Keegan_W
    Keegan_W Member Posts: 179 Red Ribbon
    edited Feb 22, 2018 12:48PM

    These are great examples - thank you so much Scott!

    I'm not sure if this is a silly question but is each elapsed_time entry for initial page load only or also for page refresh (such as generating a new chart or report on the same page with new filters)?

  • Scott Wesley
    Scott Wesley Systems Consultant & Trainer Perth, AustraliaMember Posts: 6,272 Gold Crown
    edited Feb 22, 2018 8:28PM

    I don't think it's a silly question. If I remember correctly, the page_view_type column was introduced in APEX 5.0 to help answer this question.

    pastedImage_0.png

    So you get separate records for each, and check out the request_value column to give it some more context.

This discussion has been closed.