Using srun on a SLURM cluster causes the job to exit prematurely, even with auto_requeue=False

Nobody has claimed this yet.

### Bug description

[This assertion](https://github.com/Lightning-AI/pytorch-lightning/blob/a6562b4ae7623acb87f35d49216f90aae2363b18/src/lightning/fabric/plugins/environments/slurm.py#L220) raises when inside SRUN using multiple tasks, even after doing adding `SLURMEnvironment(auto_requeue=False)`. The jobs are initially submitted through submitit-slurm.

Even after: `os.environ["SLURM_NTASKS_PER_NODE"] = os.environ["SLURM_TASKS_PER_NODE"]`, another error occurs.

Also following the advice from this thread (#6389) to do the following does not work. See also #15709.

```
from lightning.pytorch.plugins.environments import SLURMEnvironment
SLURMEnvironment.detect = lambda: False
trainer = Trainer(...)
```

There should be a simple and unambiguous way to disable _all_ special SLURM handling in lightning.

After making `_validate_srun_used`, `_validate_srun_variables`, `_is_srun_used`, `_is_slurm_interactive_mode`, all return False/early, my code runs fine! There should be a simple `disable_slurm_detection` flag, ideally also configurable as an environment variable

Thanks!

Start in src/lightning/fabric/plugins/environments/slurm.py at the _validate_srun_used assertion and review _validate_srun_variables, _is_srun_used, and _is_slurm_interactive_mode. Define a clear way to disable all special SLURM handling, including the requested configuration or environment-variable behavior, so the submitit-slurm job can run without these checks.

Written by the indexing model from the issue text.