[Core feature] Flytekit: Add support for execution name prefixes in FlyteRemote.execute()

[tomasz-sodzawiczny]

Motivation: Why do you think this is important?

Setting human-readable names might be quite for useful inspecting running executions; at the same time the behavior might be unexpected / dangerous if used naively:

Running:

ex1 = flyte_remote.execute(my_task, {"input": 1}, execution_name="foo")
ex2 = flyte_remote.execute(my_task, {"input": 2}, execution_name="foo")

Creates only 1 execution (or 0, if execution named "foo" already existed). The second set of inputs is silently ignored, and the existing execution is returned.

Goal: What should the final outcome look like, ideally?

I would suggest a way to generate a name with a human-readable prefix (with behavior similar to e.g. Kubernetes's generateName):

ex1 = flyte_remote.execute(my_task, {"input": 1}, execution_name_prefix="foo-")
ex2 = flyte_remote.execute(my_task, {"input": 2}, execution_name_prefix="foo-")

ex1.id.name # Something like "foo-41e2c72fa869"
ex2.id.name # Something like "foo-7742ba115212"

Describe alternatives you've considered

Propose: Link/Inline OR Additional context

Note: I'm happy to create a PR to flytekit if the idea seems reasonable.

Are you sure this issue hasn't been raised already?

• Yes

Have you read the Code of Conduct?

• Yes

[pingsutw]

@tomasz-sodzawiczny I think it's reasonable. Feel free to create a pr, thanks!

[troychiu]

I am working on this issue!

[troychiu]

The issue has been fixed. I think we can close it.