New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Include a Failed condition in Workloads #1982
Comments
I remember I discussed this with someone in the past. Was it you @astefanutti ? |
@alculquicondor that is likely me :) I've also touched on this with @mimowo a little while. For the UI we're building (and generally for other services that could integrate with Kueue) the Workload API provides a convenient abstraction over the different underlying workload types. There is already the It seems that logic could be encapsulated into the framework integrations, and increase the added-value of the Workload API even more. |
Everyone ok with using a Reason to indicate failure? Adding an additional condition could also be an option, but the current reason is just a placeholder. cc @tenzen-y |
Actually, this discussion happened at the KubeCon Paris 2024 with @mimowo, @astefanutti, and me :) |
/assign @trasc |
The reason and message of the finished condition should already provide this, Am I missing something? |
The purpose of this is to have the reason be "fixed", like |
In the case of Job, we have reason
So the question is whether we want to change all implementations of A separate condition could give us a standard condition plus additional granularity for failure reasons. |
Still, it might be worth it to start just with a standard |
Right, as you said previously, the current reason is just a placeholder, so this can probably be changed. As to whether introduce a separate condition, the batch Job API has one:
I don't know what drove that design decision, but if that was for "good" reasons, we could be consistent with it. |
As far as I understand, the Job API was provided before the standardized metav1.Conditions including |
I think we should just start simple. The current reason |
That sounds good to me. |
I like this idea of starting simple and using reason in the |
/assign |
What would you like to be added:
A Failed condition when a Workload finishes because the corresponding Job failed, with a reason coming from the job integration.
Just a reason in the existing Finished condition could work, although that could lead to different job integrations choosing different strings for the common idea of "failed".
Why is this needed:
Improved visibility.
Completion requirements:
This enhancement requires the following artifacts:
The artifacts should be linked in subsequent comments.
The text was updated successfully, but these errors were encountered: