Thanks for starting the discussion, Niko. In the recent days, while working on migrating K8sExecutor to use task sdk, I could clearly see some not so "nice" hardcoded executors -- CeleryK8sExecutors, SequentialExecutor (obviously), and LocalK8sExecutor!!! So, a big +1 for removing these hardcoded executors definitely for v3.
Also a +1 for adding the deprecation warnings + marking stable in providers for 2.10x or 2.11, but for core we should be looking at v3 right? Thanks & Regards, Amogh Desai On Thu, Feb 20, 2025 at 3:31 AM Oliveira, Niko <oniko...@amazon.com.invalid> wrote: > Thanks for weighing in Jed (and Jens)! And I completely agree about > marking as deprecated and stable in a 2.10.X release, I was thinking the > same today! > I'll work on a PR to make the ExecutorLoader log a deprecation warning > when it loads one of the hardcoded hybrid executors and mark multiple > executor configuration as stable. > > Cheers, > Niko > > ________________________________ > From: Jens Scheffler <j_scheff...@gmx.de.INVALID> > Sent: Wednesday, February 19, 2025 12:36:26 PM > To: dev@airflow.apache.org > Subject: RE: [EXT] [DISCUSS] Mark Multi Exec Config as stable and remove > old hardcoded hybrid execs > > CAUTION: This email originated from outside of the organization. Do not > click links or open attachments unless you can confirm the sender and know > the content is safe. > > > > AVERTISSEMENT: Ce courrier électronique provient d’un expéditeur externe. > Ne cliquez sur aucun lien et n’ouvrez aucune pièce jointe si vous ne pouvez > pas confirmer l’identité de l’expéditeur et si vous n’êtes pas certain que > le contenu ne présente aucun risque. > > > > +1 on early deprecation in next 2.10. But it is mostly about providers - > In the core we just need to clean the "if/special cases". And such > cleaning would be then the target in 3.0 > > If 1.16 of the chart closes the gap, there should be no blocker in my > view to have a 100% replacement in 3.0 w/o the need to carry legacy. > > On 19.02.25 15:54, Jed Cunningham wrote: > > I agree with Niko - chart support shouldn't be a hard line for > > experimental/stable for core features. We should get it into a released > > chart though :) It hasn't gotten enough attention from me recently. That > > said, I am close to cutting 1.16.0. > > > > I love the idea of removing hybrid support for 3. > > > > Couldn't we do the "stable" switch (removing that it's experimental from > > docs) and hybrid deprecations in the next 2.10 as well? I know we > typically > > do those in minors, but earlier the better probably at this stage. > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@airflow.apache.org > For additional commands, e-mail: dev-h...@airflow.apache.org > >
