Support asset-level timeouts

[yuhan]

What's the use case?

Users look for granular control over asset failure timings. We currently only support job-level timeouts.

Ideas of implementation

Might tackle op-level timeout at once: https://github.com/dagster-io/dagster/issues/3666

Additional information

No response

Message from the maintainers

Impacted by this issue? Give it a 👍! We factor engagement into prioritization.

[mkleinbort-ic]

Yes please - I have some assets that for whatever reason get stuck and never finish. A force terminate + retry usably fixes this.

I'd appreciate a global setting (eg. a default asset timeout)

[shanirosen-airis]

Would really appreciate as well!

[gu-xie]

+1 for me too. Got weird assets that are stuck and never finishes. Would like asset level timeout to force retries.

[AdamLenning]

+1 for me as well!

[TabLand]

FYI, I've had some success using the stopit python module. The use of an additional asset factory could be used to make this a bit more seamless until a solution is provided as part of the core offering.

https://pypi.org/project/stopit/

[MartyP233]

+1 for this, its really important when using auto materialise because you dont have a job. I have worked around this for now by using stopit around the function in the offending IOManager

[chrishiste]

+1 we had an outage because an auto materializing asset was in "preparing" mode for more than a day

[egeyumlu-cl]

+1 for me too