Do not poll for runtime job completion

[renier]

What is the expected enhancement?

For runtime jobs, job.result() checks whether the job completed or not, and keeps polling the API if it is not completed yet.

This is inefficient though, causing many requests to the API and flooding logs repetitively. In the case of a runtime job, they can run for a lot longer than the usual circuits jobs. So this polling can go on for a while.

It is possible to use the /stream endpoint which is a websocket server. If the job is still running, the websocket server will keep the stream open. Once the job has completed, the websocket server will terminate the connection. It's not possible to open a stream if the job is completed already. So if there are network errors, the client can retry connecting to the websocket until it sees the message Job is not running... along with a 302 status.

[jyu00]

@renier Does this work for programs that don't provide interim results? I guess the stream would just be left open with no data transmission?

[renier]

That’s correct.

…

-Renier

On Sep 28, 2021, at 8:09 PM, Jessie Yu ***@***.***> wrote:  @renier Does this work for programs that don't provide interim results? I guess the stream would just be left open with no data transmission? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.