Skip to content

Viewing subscriber lists takes a very long time, minutes, when lists have more than 10K subscribers

yaoquanyaoquan
edited October 2012 in Troubleshooting
Hi, I have 3 lists each with roughly 10k subscribers,and I have issues viewing subscribers list page. It takes a very long time and the server gets timed out. How can I resolve this?

Comments

  • yaoquanyaoquan
    edited October 2012
    It seems to be the count functions in includes/list/main.php. How can this be fixed?
  • yaoquanyaoquan
    Also, adding an index on subscribers.list helped to speed things up
  • nettiapinanettiapina
    edited October 2012
    I have more than 10 lists with thoussands of addresses. Kind of a way to manage 10000 message limit, and segment the postings. The subscriber lists page often takes more than 20 seconds to load, and this is on a dedicated box.

    Indeed, having DB indexes would be a step in the right direction.
  • nettiapinanettiapina
    And after doing what @yaoquan says above, it's down to under 2 sec.
  • BenBen
    @nettiapina Do you mean the mySQL query in list.php?
  • nettiapinanettiapina
    I've not modified queries in any way. Index seems to do the trick for me.
  • mattmatt
    edited October 2012
    yeah... looking at the code.... it gets * for each list and then loops thru each one and gets count data for unsubscribes / bounces etc....

    when we have a large amount of data... it is one of those places you just hate to go... ( damn... I gotta go back into the list page )....

    WHAT IF

    you just show a the list of lists... and if we click the INFO button... then we can get updated info on that.... via AJAX

    AND

    what if you keep that data TOTALED in the LIST table and update if when an event happens to that list...

    LIKE

    someone unsubscribes from an email in that list... you update the LIST row in the LIST table ...

    WHAM

    that would be one of the fastest pages in the whole system...

    just a thought....

    m
  • DrCocktorDrCocktor
    Alternatively, you could simply 'SELECT COUNT(id) AS count FROM ...' and then simply have a single count value returned, an action which the database can perform much faster than actually querying and returning an entire dataset and then having to walk over that dataset to count the number of rows.
  • BenBen
    @nettiapina @matt @DrCocktor Thanks for the suggestions, will definitely change this.
This discussion has been closed.

© Sendy.co 2024