You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@beam.apache.org by Artur Khanin <ar...@akvelon.com> on 2022/03/21 21:18:36 UTC
Re: [EXTERNAL] Collect And Deploy Playground Examples - GH action is failing / flaky
Hi all,
Thank you for bringing this up! I filed a ticket BEAM-14140<https://issues.apache.org/jira/browse/BEAM-14140>, and we started working on this.
We want to approach it by gathering the list of changed examples and if there is any — check them that they aren’t broken. Does it seem for you the right way to solve it?
Thanks,
Artur Khanin
On 21 Mar 2022, at 23:57, Ahmet Altay <al...@google.com>> wrote:
On Mon, Mar 21, 2022 at 12:28 PM Robert Bradshaw <ro...@google.com>> wrote:
On Mon, Mar 21, 2022 at 11:56 AM Ahmet Altay <al...@google.com>> wrote:
>
> On Mon, Mar 21, 2022 at 8:57 AM Pablo Estrada <pa...@google.com>> wrote:
>>
>> Artur, Ilya, do you know why this action is perma-red?
>
> Thank you for looking folks. And if there is not one, could you also please file a jira issue?
>
>> Best
>> -P.
>>
>> On Mon, Mar 21, 2022 at 8:39 AM Jack McCluskey <jr...@google.com>> wrote:
>>>
>>> Does this action need to run on every single PR? As far as I can tell it isn't saying anything about most PRs being run against it, pass or fail.
>
> I agree. Probably does not need to run on every PR. It could be a scheduled action instead.
This would drastically reduce its usefulness in determining
whether/which PR brakes it. I would remove it as a precommit iff it's
the kind of thing that we would not expect the typical PR to influence
(e.g. could it be triggered on changes influencing a subdirectory).
Of course the short term option of moving/disabling it until it
provides an actual signal (i.e. it's green unless there are actual
regressions) makes sense.
My _understanding_: This action is used for detecting any new examples added to the repo and tags those for inclusion in the beam playground. If that is correct a post commit might be a better option. A scheduled GH action would be similar to a post commit.
>>> On Fri, Mar 18, 2022 at 11:06 PM Ahmet Altay <al...@google.com>> wrote:
>>>>
>>>> Hi all,
>>>>
>>>> This GH action seems to be recently added and permanently failing [1]. Is there a JIRA for addressing this? I assume not, but just checking, should it be blocking merges?
>>>>
>>>> Thank you!
>>>> Ahmet
>>>>
>>>> [1] https://github.com/apache/beam/actions/workflows/playground_deploy_examples.yml
Re: [EXTERNAL] Collect And Deploy Playground Examples - GH action is failing / flaky
Posted by Ahmet Altay <al...@google.com>.
Thank you Luke! I agree with disabling manually.
On Tue, Mar 22, 2022 at 4:07 PM Luke Cwik <lc...@google.com> wrote:
> I have manually disabled the GH Action from the UI as there have not been
> any successful runs of the action in the couple of months it has been
> active.
>
> I filed https://issues.apache.org/jira/browse/BEAM-14154 as that seems
> different then BEAM-14140 which is about filtering when it is run.
>
> On Mon, Mar 21, 2022 at 5:59 PM Ahmet Altay <al...@google.com> wrote:
>
>>
>>
>> On Mon, Mar 21, 2022 at 3:53 PM Pablo Estrada <pa...@google.com> wrote:
>>
>>> These PRs can just be ignored. This workflow is related to Beam
>>> Playground - as long as a PR is not related to the Beam Playground, you can
>>> ignore the failed workflow.
>>>
>>> On Mon, Mar 21, 2022 at 3:43 PM Lara Schmidt <la...@google.com>
>>> wrote:
>>>
>>>> For PRs that are failing with this, what's the recommended path?
>>>> Should we wait for a fix? Remove it until one can be provided? Or just
>>>> ignore it?
>>>>
>>>> On Mon, Mar 21, 2022 at 2:18 PM Artur Khanin <ar...@akvelon.com>
>>>> wrote:
>>>>
>>>>> Hi all,
>>>>>
>>>>> Thank you for bringing this up! I filed a ticket BEAM-14140
>>>>> <https://issues.apache.org/jira/browse/BEAM-14140>, and we started
>>>>> working on this.
>>>>>
>>>>
>> Thank you.
>>
>>
>>> We want to approach it by gathering the list of changed examples and if
>>>>> there is any — check them that they aren’t broken. Does it seem for you
>>>>> the right way to solve it?
>>>>>
>>>>
>> That sounds reasonable to me.
>>
>> My only request would be to please make the action as stable as possible.
>> This is true for all precommits. We would like it to catch broken examples,
>> with a minimal amount of flakes.
>>
>>
>>
>>>
>>>>> Thanks,
>>>>> Artur Khanin
>>>>>
>>>>> On 21 Mar 2022, at 23:57, Ahmet Altay <al...@google.com> wrote:
>>>>>
>>>>>
>>>>>
>>>>> On Mon, Mar 21, 2022 at 12:28 PM Robert Bradshaw <ro...@google.com>
>>>>> wrote:
>>>>>
>>>>>> On Mon, Mar 21, 2022 at 11:56 AM Ahmet Altay <al...@google.com>
>>>>>> wrote:
>>>>>> >
>>>>>> > On Mon, Mar 21, 2022 at 8:57 AM Pablo Estrada <pa...@google.com>
>>>>>> wrote:
>>>>>> >>
>>>>>> >> Artur, Ilya, do you know why this action is perma-red?
>>>>>> >
>>>>>> > Thank you for looking folks. And if there is not one, could you
>>>>>> also please file a jira issue?
>>>>>> >
>>>>>> >> Best
>>>>>> >> -P.
>>>>>> >>
>>>>>> >> On Mon, Mar 21, 2022 at 8:39 AM Jack McCluskey <
>>>>>> jrmccluskey@google.com> wrote:
>>>>>> >>>
>>>>>> >>> Does this action need to run on every single PR? As far as I can
>>>>>> tell it isn't saying anything about most PRs being run against it, pass or
>>>>>> fail.
>>>>>> >
>>>>>> > I agree. Probably does not need to run on every PR. It could be a
>>>>>> scheduled action instead.
>>>>>>
>>>>>> This would drastically reduce its usefulness in determining
>>>>>> whether/which PR brakes it. I would remove it as a precommit iff it's
>>>>>> the kind of thing that we would not expect the typical PR to influence
>>>>>> (e.g. could it be triggered on changes influencing a subdirectory).
>>>>>>
>>>>>> Of course the short term option of moving/disabling it until it
>>>>>> provides an actual signal (i.e. it's green unless there are actual
>>>>>> regressions) makes sense.
>>>>>>
>>>>>
>>>>> My _understanding_: This action is used for detecting any new examples
>>>>> added to the repo and tags those for inclusion in the beam playground. If
>>>>> that is correct a post commit might be a better option. A scheduled GH
>>>>> action would be similar to a post commit.
>>>>>
>>>>>
>>>>>>
>>>>>> >>> On Fri, Mar 18, 2022 at 11:06 PM Ahmet Altay <al...@google.com>
>>>>>> wrote:
>>>>>> >>>>
>>>>>> >>>> Hi all,
>>>>>> >>>>
>>>>>> >>>> This GH action seems to be recently added and permanently
>>>>>> failing [1]. Is there a JIRA for addressing this? I assume not, but just
>>>>>> checking, should it be blocking merges?
>>>>>> >>>>
>>>>>> >>>> Thank you!
>>>>>> >>>> Ahmet
>>>>>> >>>>
>>>>>> >>>> [1]
>>>>>> https://github.com/apache/beam/actions/workflows/playground_deploy_examples.yml
>>>>>
>>>>>
>>>>>
Re: [EXTERNAL] Collect And Deploy Playground Examples - GH action is failing / flaky
Posted by Luke Cwik <lc...@google.com>.
I have manually disabled the GH Action from the UI as there have not been
any successful runs of the action in the couple of months it has been
active.
I filed https://issues.apache.org/jira/browse/BEAM-14154 as that seems
different then BEAM-14140 which is about filtering when it is run.
On Mon, Mar 21, 2022 at 5:59 PM Ahmet Altay <al...@google.com> wrote:
>
>
> On Mon, Mar 21, 2022 at 3:53 PM Pablo Estrada <pa...@google.com> wrote:
>
>> These PRs can just be ignored. This workflow is related to Beam
>> Playground - as long as a PR is not related to the Beam Playground, you can
>> ignore the failed workflow.
>>
>> On Mon, Mar 21, 2022 at 3:43 PM Lara Schmidt <la...@google.com>
>> wrote:
>>
>>> For PRs that are failing with this, what's the recommended path?
>>> Should we wait for a fix? Remove it until one can be provided? Or just
>>> ignore it?
>>>
>>> On Mon, Mar 21, 2022 at 2:18 PM Artur Khanin <ar...@akvelon.com>
>>> wrote:
>>>
>>>> Hi all,
>>>>
>>>> Thank you for bringing this up! I filed a ticket BEAM-14140
>>>> <https://issues.apache.org/jira/browse/BEAM-14140>, and we started
>>>> working on this.
>>>>
>>>
> Thank you.
>
>
>> We want to approach it by gathering the list of changed examples and if
>>>> there is any — check them that they aren’t broken. Does it seem for you
>>>> the right way to solve it?
>>>>
>>>
> That sounds reasonable to me.
>
> My only request would be to please make the action as stable as possible.
> This is true for all precommits. We would like it to catch broken examples,
> with a minimal amount of flakes.
>
>
>
>>
>>>> Thanks,
>>>> Artur Khanin
>>>>
>>>> On 21 Mar 2022, at 23:57, Ahmet Altay <al...@google.com> wrote:
>>>>
>>>>
>>>>
>>>> On Mon, Mar 21, 2022 at 12:28 PM Robert Bradshaw <ro...@google.com>
>>>> wrote:
>>>>
>>>>> On Mon, Mar 21, 2022 at 11:56 AM Ahmet Altay <al...@google.com> wrote:
>>>>> >
>>>>> > On Mon, Mar 21, 2022 at 8:57 AM Pablo Estrada <pa...@google.com>
>>>>> wrote:
>>>>> >>
>>>>> >> Artur, Ilya, do you know why this action is perma-red?
>>>>> >
>>>>> > Thank you for looking folks. And if there is not one, could you also
>>>>> please file a jira issue?
>>>>> >
>>>>> >> Best
>>>>> >> -P.
>>>>> >>
>>>>> >> On Mon, Mar 21, 2022 at 8:39 AM Jack McCluskey <
>>>>> jrmccluskey@google.com> wrote:
>>>>> >>>
>>>>> >>> Does this action need to run on every single PR? As far as I can
>>>>> tell it isn't saying anything about most PRs being run against it, pass or
>>>>> fail.
>>>>> >
>>>>> > I agree. Probably does not need to run on every PR. It could be a
>>>>> scheduled action instead.
>>>>>
>>>>> This would drastically reduce its usefulness in determining
>>>>> whether/which PR brakes it. I would remove it as a precommit iff it's
>>>>> the kind of thing that we would not expect the typical PR to influence
>>>>> (e.g. could it be triggered on changes influencing a subdirectory).
>>>>>
>>>>> Of course the short term option of moving/disabling it until it
>>>>> provides an actual signal (i.e. it's green unless there are actual
>>>>> regressions) makes sense.
>>>>>
>>>>
>>>> My _understanding_: This action is used for detecting any new examples
>>>> added to the repo and tags those for inclusion in the beam playground. If
>>>> that is correct a post commit might be a better option. A scheduled GH
>>>> action would be similar to a post commit.
>>>>
>>>>
>>>>>
>>>>> >>> On Fri, Mar 18, 2022 at 11:06 PM Ahmet Altay <al...@google.com>
>>>>> wrote:
>>>>> >>>>
>>>>> >>>> Hi all,
>>>>> >>>>
>>>>> >>>> This GH action seems to be recently added and permanently failing
>>>>> [1]. Is there a JIRA for addressing this? I assume not, but just checking,
>>>>> should it be blocking merges?
>>>>> >>>>
>>>>> >>>> Thank you!
>>>>> >>>> Ahmet
>>>>> >>>>
>>>>> >>>> [1]
>>>>> https://github.com/apache/beam/actions/workflows/playground_deploy_examples.yml
>>>>
>>>>
>>>>
Re: [EXTERNAL] Collect And Deploy Playground Examples - GH action is failing / flaky
Posted by Ahmet Altay <al...@google.com>.
On Mon, Mar 21, 2022 at 3:53 PM Pablo Estrada <pa...@google.com> wrote:
> These PRs can just be ignored. This workflow is related to Beam Playground
> - as long as a PR is not related to the Beam Playground, you can ignore the
> failed workflow.
>
> On Mon, Mar 21, 2022 at 3:43 PM Lara Schmidt <la...@google.com>
> wrote:
>
>> For PRs that are failing with this, what's the recommended path?
>> Should we wait for a fix? Remove it until one can be provided? Or just
>> ignore it?
>>
>> On Mon, Mar 21, 2022 at 2:18 PM Artur Khanin <ar...@akvelon.com>
>> wrote:
>>
>>> Hi all,
>>>
>>> Thank you for bringing this up! I filed a ticket BEAM-14140
>>> <https://issues.apache.org/jira/browse/BEAM-14140>, and we started
>>> working on this.
>>>
>>
Thank you.
> We want to approach it by gathering the list of changed examples and if
>>> there is any — check them that they aren’t broken. Does it seem for you
>>> the right way to solve it?
>>>
>>
That sounds reasonable to me.
My only request would be to please make the action as stable as possible.
This is true for all precommits. We would like it to catch broken examples,
with a minimal amount of flakes.
>
>>> Thanks,
>>> Artur Khanin
>>>
>>> On 21 Mar 2022, at 23:57, Ahmet Altay <al...@google.com> wrote:
>>>
>>>
>>>
>>> On Mon, Mar 21, 2022 at 12:28 PM Robert Bradshaw <ro...@google.com>
>>> wrote:
>>>
>>>> On Mon, Mar 21, 2022 at 11:56 AM Ahmet Altay <al...@google.com> wrote:
>>>> >
>>>> > On Mon, Mar 21, 2022 at 8:57 AM Pablo Estrada <pa...@google.com>
>>>> wrote:
>>>> >>
>>>> >> Artur, Ilya, do you know why this action is perma-red?
>>>> >
>>>> > Thank you for looking folks. And if there is not one, could you also
>>>> please file a jira issue?
>>>> >
>>>> >> Best
>>>> >> -P.
>>>> >>
>>>> >> On Mon, Mar 21, 2022 at 8:39 AM Jack McCluskey <
>>>> jrmccluskey@google.com> wrote:
>>>> >>>
>>>> >>> Does this action need to run on every single PR? As far as I can
>>>> tell it isn't saying anything about most PRs being run against it, pass or
>>>> fail.
>>>> >
>>>> > I agree. Probably does not need to run on every PR. It could be a
>>>> scheduled action instead.
>>>>
>>>> This would drastically reduce its usefulness in determining
>>>> whether/which PR brakes it. I would remove it as a precommit iff it's
>>>> the kind of thing that we would not expect the typical PR to influence
>>>> (e.g. could it be triggered on changes influencing a subdirectory).
>>>>
>>>> Of course the short term option of moving/disabling it until it
>>>> provides an actual signal (i.e. it's green unless there are actual
>>>> regressions) makes sense.
>>>>
>>>
>>> My _understanding_: This action is used for detecting any new examples
>>> added to the repo and tags those for inclusion in the beam playground. If
>>> that is correct a post commit might be a better option. A scheduled GH
>>> action would be similar to a post commit.
>>>
>>>
>>>>
>>>> >>> On Fri, Mar 18, 2022 at 11:06 PM Ahmet Altay <al...@google.com>
>>>> wrote:
>>>> >>>>
>>>> >>>> Hi all,
>>>> >>>>
>>>> >>>> This GH action seems to be recently added and permanently failing
>>>> [1]. Is there a JIRA for addressing this? I assume not, but just checking,
>>>> should it be blocking merges?
>>>> >>>>
>>>> >>>> Thank you!
>>>> >>>> Ahmet
>>>> >>>>
>>>> >>>> [1]
>>>> https://github.com/apache/beam/actions/workflows/playground_deploy_examples.yml
>>>
>>>
>>>
Re: [EXTERNAL] Collect And Deploy Playground Examples - GH action is failing / flaky
Posted by Pablo Estrada <pa...@google.com>.
These PRs can just be ignored. This workflow is related to Beam Playground
- as long as a PR is not related to the Beam Playground, you can ignore the
failed workflow.
On Mon, Mar 21, 2022 at 3:43 PM Lara Schmidt <la...@google.com> wrote:
> For PRs that are failing with this, what's the recommended path? Should we
> wait for a fix? Remove it until one can be provided? Or just ignore it?
>
> On Mon, Mar 21, 2022 at 2:18 PM Artur Khanin <ar...@akvelon.com>
> wrote:
>
>> Hi all,
>>
>> Thank you for bringing this up! I filed a ticket BEAM-14140
>> <https://issues.apache.org/jira/browse/BEAM-14140>, and we started
>> working on this.
>> We want to approach it by gathering the list of changed examples and if
>> there is any — check them that they aren’t broken. Does it seem for you
>> the right way to solve it?
>>
>> Thanks,
>> Artur Khanin
>>
>> On 21 Mar 2022, at 23:57, Ahmet Altay <al...@google.com> wrote:
>>
>>
>>
>> On Mon, Mar 21, 2022 at 12:28 PM Robert Bradshaw <ro...@google.com>
>> wrote:
>>
>>> On Mon, Mar 21, 2022 at 11:56 AM Ahmet Altay <al...@google.com> wrote:
>>> >
>>> > On Mon, Mar 21, 2022 at 8:57 AM Pablo Estrada <pa...@google.com>
>>> wrote:
>>> >>
>>> >> Artur, Ilya, do you know why this action is perma-red?
>>> >
>>> > Thank you for looking folks. And if there is not one, could you also
>>> please file a jira issue?
>>> >
>>> >> Best
>>> >> -P.
>>> >>
>>> >> On Mon, Mar 21, 2022 at 8:39 AM Jack McCluskey <
>>> jrmccluskey@google.com> wrote:
>>> >>>
>>> >>> Does this action need to run on every single PR? As far as I can
>>> tell it isn't saying anything about most PRs being run against it, pass or
>>> fail.
>>> >
>>> > I agree. Probably does not need to run on every PR. It could be a
>>> scheduled action instead.
>>>
>>> This would drastically reduce its usefulness in determining
>>> whether/which PR brakes it. I would remove it as a precommit iff it's
>>> the kind of thing that we would not expect the typical PR to influence
>>> (e.g. could it be triggered on changes influencing a subdirectory).
>>>
>>> Of course the short term option of moving/disabling it until it
>>> provides an actual signal (i.e. it's green unless there are actual
>>> regressions) makes sense.
>>>
>>
>> My _understanding_: This action is used for detecting any new examples
>> added to the repo and tags those for inclusion in the beam playground. If
>> that is correct a post commit might be a better option. A scheduled GH
>> action would be similar to a post commit.
>>
>>
>>>
>>> >>> On Fri, Mar 18, 2022 at 11:06 PM Ahmet Altay <al...@google.com>
>>> wrote:
>>> >>>>
>>> >>>> Hi all,
>>> >>>>
>>> >>>> This GH action seems to be recently added and permanently failing
>>> [1]. Is there a JIRA for addressing this? I assume not, but just checking,
>>> should it be blocking merges?
>>> >>>>
>>> >>>> Thank you!
>>> >>>> Ahmet
>>> >>>>
>>> >>>> [1]
>>> https://github.com/apache/beam/actions/workflows/playground_deploy_examples.yml
>>
>>
>>
Re: [EXTERNAL] Collect And Deploy Playground Examples - GH action is failing / flaky
Posted by Lara Schmidt <la...@google.com>.
For PRs that are failing with this, what's the recommended path? Should we
wait for a fix? Remove it until one can be provided? Or just ignore it?
On Mon, Mar 21, 2022 at 2:18 PM Artur Khanin <ar...@akvelon.com>
wrote:
> Hi all,
>
> Thank you for bringing this up! I filed a ticket BEAM-14140
> <https://issues.apache.org/jira/browse/BEAM-14140>, and we started
> working on this.
> We want to approach it by gathering the list of changed examples and if
> there is any — check them that they aren’t broken. Does it seem for you
> the right way to solve it?
>
> Thanks,
> Artur Khanin
>
> On 21 Mar 2022, at 23:57, Ahmet Altay <al...@google.com> wrote:
>
>
>
> On Mon, Mar 21, 2022 at 12:28 PM Robert Bradshaw <ro...@google.com>
> wrote:
>
>> On Mon, Mar 21, 2022 at 11:56 AM Ahmet Altay <al...@google.com> wrote:
>> >
>> > On Mon, Mar 21, 2022 at 8:57 AM Pablo Estrada <pa...@google.com>
>> wrote:
>> >>
>> >> Artur, Ilya, do you know why this action is perma-red?
>> >
>> > Thank you for looking folks. And if there is not one, could you also
>> please file a jira issue?
>> >
>> >> Best
>> >> -P.
>> >>
>> >> On Mon, Mar 21, 2022 at 8:39 AM Jack McCluskey <jr...@google.com>
>> wrote:
>> >>>
>> >>> Does this action need to run on every single PR? As far as I can tell
>> it isn't saying anything about most PRs being run against it, pass or fail.
>> >
>> > I agree. Probably does not need to run on every PR. It could be a
>> scheduled action instead.
>>
>> This would drastically reduce its usefulness in determining
>> whether/which PR brakes it. I would remove it as a precommit iff it's
>> the kind of thing that we would not expect the typical PR to influence
>> (e.g. could it be triggered on changes influencing a subdirectory).
>>
>> Of course the short term option of moving/disabling it until it
>> provides an actual signal (i.e. it's green unless there are actual
>> regressions) makes sense.
>>
>
> My _understanding_: This action is used for detecting any new examples
> added to the repo and tags those for inclusion in the beam playground. If
> that is correct a post commit might be a better option. A scheduled GH
> action would be similar to a post commit.
>
>
>>
>> >>> On Fri, Mar 18, 2022 at 11:06 PM Ahmet Altay <al...@google.com>
>> wrote:
>> >>>>
>> >>>> Hi all,
>> >>>>
>> >>>> This GH action seems to be recently added and permanently failing
>> [1]. Is there a JIRA for addressing this? I assume not, but just checking,
>> should it be blocking merges?
>> >>>>
>> >>>> Thank you!
>> >>>> Ahmet
>> >>>>
>> >>>> [1]
>> https://github.com/apache/beam/actions/workflows/playground_deploy_examples.yml
>
>
>