fix: Don't retry indexing if key size is too long

[chubei]

The key size error used to lead to dead loop and indexing cannot make progress.

Panic message is like

ERROR panicked at dozer-cache/src/cache/lmdb/indexing.rs:226:21:
Secondary index key is too long. This usually happens with `String` fields. Try to [skip](https://getdozer.io/docs/configuration/api-endpoints#indexes) creating secondary index SortedInverted([5]).

[Jesse-Bakker]

We have both String and Text types. Should we have a size restriction on String and restrict creating SortedInverted indexes to String?

[chubei]

We have both String and Text types. Should we have a size restriction on String and restrict creating SortedInverted indexes to String?

That would be the way to go.