BUG: prevent overflow in npy_datetimestruct_to_datetime

[WillAyd]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

We have a lot of multiplication in npy_datetimestruct_to_datetime that can overflow:

pandas/pandas/_libs/src/vendored/numpy/datetime/np_datetime.c

Line 308 in 79067a7

npy_datetime npy_datetimestruct_to_datetime(NPY_DATETIMEUNIT base,

UBSAN detects this from running test_sql.py (first error only shown):

/pandas/_libs/src/vendored/numpy/datetime/np_datetime.c:359:33: runtime error: signed integer overflow: 253370764800000000 * 1000 cannot be represented in type 'long int'

What will be tricky about this function is it has an integral return value, where all values are technically valid. Ideally we have some way to avoid the overflow and signal an error via CPython

Issue Description

see above

Expected Behavior

see above

Installed Versions

main

[tpcgold]

running also into this issue since today. getting data from bigquery to dataframe and then pandas fails to convert it to json

  File "<ipython-input-2-fa9883de5fe4>", line 82, in my_function
    rows = df.to_json(orient="records")
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/PycharmProjects/data-etl/venv/lib/python3.11/site-packages/pandas/util/_decorators.py", line 333, in wrapper
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/PycharmProjects/data-etl/venv/lib/python3.11/site-packages/pandas/core/generic.py", line 2699, in to_json
    return json.to_json(
           ^^^^^^^^^^^^^
  File "/PycharmProjects/data-etl/venv/lib/python3.11/site-packages/pandas/io/json/_json.py", line 210, in to_json
    ).write()
      ^^^^^^^
  File "/PycharmProjects/data-etl/venv/lib/python3.11/site-packages/pandas/io/json/_json.py", line 263, in write
    return ujson_dumps(
           ^^^^^^^^^^^^
OverflowError: Overflow occurred in npy_datetimestruct_to_datetime

[moble]

Maybe, like #57349, this is also closed by #55901?