Job status - average time to create or update many users

2 Comments

  • Joseph May

    Hi Nicolas-
    Good question; I am unsure as to whether or not there is any standard here - I will ask some colleagues for input. That being said, job status data is available only for an hour after a given job is created, per our documentation. My best suggestion, barring any specific averages, would be to create some test resources and query at intervals in line with your goals. In your example of 100, I would start with querying at 5 minutes after and increment up from there.
    If I am able to gather any more specifics I will let you know, but bear in mind that the user object will vary from account to account, and even an average may not be telling.

    0
  • Joseph May

    Hi Nicolas-

    I spoke with some colleagues who provided some detailed feedback.

    The length of the job depends on “number of items” * “average save time per item”
    • number of items = dependent on the request
    • average save time = varies per user. Some customers save resources in a second. Some take 15 seconds. The variance depends on complexity

    Possible techniques for optimization:

    1. Create an estimate for how long each item takes to save & multiple it by the number of items. Make it a little early by some coefficient, and then poll each second afterwards, e.g., If a user takes 2 seconds, and you’re updating 10 of them, maybe start polling at the 15 second mark, and then again every second afterwards. This should help determine a baseline for you to work with.

    2. Be adaptive. The status response provides how many items have been worked through. Poll it a couple times and create a regression for that given request.

    0

Please sign in to leave a comment.

Powered by Zendesk