Skip to content

Asynchronous workflows retry policies

If a workflow is published in asynchronous mode, it has an automatically inserted additional block that manages the request queue.
Compared to a synchronous workflow, this allows for retrying analyses if, for some reason, the workflow doesn't respond. There is therefore no need to touch the client software to manage analysis attempts, because the workflow manages them on its own in a transparent way.
The following parameters, to be set during publication, are the retry policies.

  • Attempts: number of execution attempts. If the workflow does not respond, as many attempts as this number plus one are made. The attempts are spaced out by an interval that is configured with the two parameters below.
  • Delay: basic delay in seconds for calculating the interval between attempts in seconds.
    The interval between the first and second attempt is:

    Delay + (Delay X Multiplier)

    The interval from the second attempt onwards is:

    Previous interval + (Delay X Multiplier)
  • Multiplier: multiplier used to calculate the interval between attempts (see above).

For example, if:

Attempts = 5
Delay = 2s
Multiplier = 1

a maximum of six attempts will be made with the following intervals between them:

  1. 2s X 1 = 2s
  2. 2s + (2s X 1) = 4s
  3. 4s + (2s X 1) = 6s
  4. 6s + (2x X 1) = 8s
  5. 8s + (2s X 1) = 10s