Tool generates incomplete nested collection when quota is violated

**Describe the bug**

I have a single job creating list and list:list collection(s) using dataset discovery. The job creates more data than the user's quota. I tested separately the generation of list and list:list:

- some of the members of the list:list collections stay empty and all collections and contained datasets (if present) are shown green.
- For list collections all collection elements seemed fine (quota was violated after the job) .

For list:list there should be some indication of error or all datasets should be visible (for consistence with the `list` case I would prefer to keep the data).

Edit: the above was with metadata strategy: directory. I also tried with extended.

I have to add the the used tool generates a dataset and list or list:list collections (which collections can be chosen by the user) .. the tool is NCBI datasets. After the job finishes:

- the collections end up empty in the history
- datasets remains yellow (I waited 12h)

**Galaxy Version and/or server at which you observed the bug**
Galaxy Version: 22.05
Commit: b709a19c5edbb1015145aa7aad074ee17cd9724b

**To Reproduce**

- use a tool creating a large collection (I use NCBI Datasets from IUC)
- create more data that your quota

**Expected behavior**

Either

1. Error should be indicated.
2. Reset job and set to paused .. indicating that quota was violated.
3. Add all data to the collections (which means that the user would get to >100% quota usage)

For 1. and 2 the job should never be successful.

Reproduce the issue with the NCBI Datasets tool by creating a nested collection that exceeds the user quota, then inspect the resulting history collections and datasets. Determine whether quota violations leave the job failed or paused with a visible error, or preserve all collection data, and verify that the job is not reported successful in the chosen behavior.

Written by the indexing model from the issue text.