great-egg-84692
08/26/2025, 5:02 PMancient-application-36103
08/26/2025, 5:33 PMgreat-egg-84692
08/26/2025, 5:50 PMgreat-egg-84692
08/26/2025, 7:02 PMancient-application-36103
08/26/2025, 7:02 PMancient-application-36103
08/26/2025, 7:02 PMancient-application-36103
08/26/2025, 7:05 PMreplace
as a policy? i wonder, if there was support for controlling workflow concurrency - would that also solve for this?great-egg-84692
08/26/2025, 7:06 PMreplace
I'm mainly looking for using Forbid
.
if there was support for controlling workflow concurrency - would that also solve for this?sounds there is an alternative I am not aware of
ancient-application-36103
08/26/2025, 7:12 PMgreat-egg-84692
08/26/2025, 7:14 PMancient-application-36103
08/26/2025, 7:18 PMancient-application-36103
08/26/2025, 7:21 PMancient-application-36103
08/26/2025, 7:32 PMmutex
would be that it will work across all triggering primitives - cron, manual, eventsgreat-egg-84692
08/26/2025, 7:32 PMancient-application-36103
08/26/2025, 7:44 PMmutex
implementation just adds a field to the workflow templateancient-application-36103
08/26/2025, 7:46 PMgreat-egg-84692
09/04/2025, 5:30 PMgreat-egg-84692
09/04/2025, 5:31 PMalso, are you looking for a queuing semantic or an explicit deny?but for my use case, I feel an explicit deny is closer to what I want.
thankful-ambulance-42457
09/09/2025, 1:24 PM@schedule
decorator if it is to be generalised for all workflows on argo. a separate decorator could be a better fit.
concurrencyPolicy:
⢠also supports `Replace`which might have its uses
⢠not necessarily simpler to define a policy name and ties this directly to argo verbiage
⢠option makes sense in the @schedule
decorator from Argo Workflows perspective
all things considered, I would be leaning more towards a general solution that sets an upper bound on concurrent executions (semaphore)great-egg-84692
09/09/2025, 1:46 PMthankful-ambulance-42457
09/09/2025, 2:26 PM@schedule
is a great fit for defining a generic concurrency limit without setting a schedule š
Glad to hear about the fork, that's definitely the quickest option forward for the time being š
Just for future reference, implementing something like @schedule(concurrency_limit=n)
should address your use case as well with n=1
, correct?great-egg-84692
09/09/2025, 2:29 PMJust for future reference, implementing something likeyeah correct.should address your use case as well with@schedule(concurrency_limit=n)
, correct?n=1
great-egg-84692
09/09/2025, 2:31 PM@schedule
is a great fit for defining a generic concurrency limit without setting a schedule
yeah, this makes sense, but I also wonder why argo has both Semaphore/mutex
and concurrencyPolicy
, it feels concurrencyPolicy
(tho simpler) isn't needed given Semaphore/mutex
exists. Do you have any insights?thankful-ambulance-42457
09/09/2025, 3:17 PMReplace
which can't be implemented with a simple semaphore