Thanks Caizhi, our use case is for testing - before running a production job, we want to “dry run” it for a while and output the result to a temp location where we can verify the output. We are running a streaming job but the use case sounds more like a batch job.
On Mon, Oct 11, 2021 at 7:34 PM Caizhi Weng <tsreape...@gmail.com> wrote: > Hi! > > There is currently no such setting. You need to rely on an external system > to read the execution time (from Flink's job information, see [1]) and > cancel the job once it exceeds the time limit. > > Could you elaborate more on your use case? Are you running a streaming job > or a batch job? For streaming jobs it is rare to have a time limit. > > [1] > https://nightlies.apache.org/flink/flink-docs-release-1.13/docs/ops/rest_api/#jobs-jobid > > Sharon Xie <sharon.xie...@gmail.com> 于2021年10月12日周二 上午3:42写道: > >> Hi there, >> >> We have a use case where we want to terminate a job when a time limit >> is reached. Is there a Flink setting that we can use for this use case? >> >> >> Thanks, >> Sharon >> >