Reason for "dataclip not saved" incorrect or misleading

It seems the "retry run from step" tooltip (when input data clip was not returned) is misleading.

Update: The new input component may not handle this at all: https://www.loom.com/share/5c65264a08024a56b8ceff6b167e35a8

When the data clip size is larger than the limit parsed into run options, the worker does not send the data clip back but uses it to process the next step without breaking the run. This is great, but the issue is that when viewing the run history, the messaging is misleading for users. It currently reads: "This work order cannot be rerun since no input data has been stored due to the data retention policy set in the project.
Go to data storage settings"_.

We should change the tooltip message to read: "The input data clip for this step was to be stored. Check the run logs for more information."

PS: We have fixed a related issue for steps that were not included in a run here. So it might be helpful to note that this is not a one-fix for all.

We should ensure that all of the various reasons that a dataclip is NOT stored are accounted for and tested.

• too big?
• slow network - it's loading but not yet loaded
• wasn't sent by worker? (never received?)
• deleted because it expired (input/output retention period)
• deleled because it's disabled (zero-persistence for input/output)
• others??? (can you help figure out which other reasons we might have?)

Start at the run-history "retry run from step" tooltip and compare its message with the run logs and data storage settings. Trace the listed reasons a dataclip may not be stored, including size, delayed or missing worker responses, expiry, and disabled persistence. Done means each reason has accurate messaging and is covered by tests.