Rally Software

  • 1.  REST Timeout and Pagesize inconsistency

    Posted Jul 20, 2017 04:23 PM

    When executing a REST Get, is there any way to set the fetch to exclude certain fields rather than having to explicitly list everything except the fields like APIMajor, APIMinor, etc?

     

    I've experienced inconsistent performance over the last month.  Last week I successfully extracted the entire Task object with a page size of 2000, today it failed for 2000, 1000, and finally ran successfully for 500.  Two weeks ago I could extract HierarchicalRequirement with a page size of 250, but just now it failed with a page size of 100 after pulling half of the records.

     

    Often, a pagesize of 2000 wont even successfully return results in the graph view to the API documentation interface or when extracting a single row from the header for total row count

     

    Adjusting the timeout seems to have little impact, yet the failure indicates that the extract is timing out.

     

    Any suggestions?



  • 2.  Re: REST Timeout and Pagesize inconsistency
    Best Answer

    Posted Jul 20, 2017 06:01 PM

    Hi chance.k.barkley,

     

    Pat Cousins logged a ticket with support for what sounds like the same issue in your subscription. We have logged  defect DE35538 with our engineering team to track the issue.

     

    For your question about fetch, you must explicitly list out the attributes in your request or set fetch=true to return everything.

     

    Thanks for reporting this to us!

     

    -Sean Davis



  • 3.  Re: REST Timeout and Pagesize inconsistency

    Posted Oct 04, 2017 03:10 PM

    I am having the same issue. Where can I find status/notes for DE35538? What is the status of DE35538? Is there a work-around? Or will this be addressed in an upcoming release?



  • 4.  Re: REST Timeout and Pagesize inconsistency

    Posted Oct 04, 2017 03:50 PM

    Hi Joe,

     

    Unfortunately we do not have a public place to view defect status. This particular defect was resolved and was specific to certain request types. I would suggest logging a support ticket with the following details:

     

    1. The UserID used for the slow request(s).

    2. The date/time and timezone the request(s) were run.

    3. The full request message(s).

    4. The full response message(s).

     

    -Sean Davis