Evacuate chunk on networking issues need limits

### Describe the bug

We output to multiple locations, one of the locations was down for a long period of time and fluent began evacuating chunks to the root_dir until it filled up the disk. It would be good to add an option to limit the size of the evacuated chunks or disable the functionality

### To Reproduce

Make an output unreachable/timeout and let the logs flow until root_dir is full.

### Expected behavior

Evacuated chunks should have a limit to prevent filling up the disk or the feature should be able to be disabled. Or both

### Your Environment

```markdown
- Fluentd version: 1.19.0
- Package version:
- Operating system:
- Kernel version:
```

### Your Configuration

```apache
Any config with an output that can be unavailable/unreachable
```

### Your Error Log

```shell
2026-05-04 15:15:29 +0000 [warn]: #0 chunk files are evacuated to /tmp/fluent/buffer/object_570. chunk_id="650ff4c25bf36ff8208adc20ecb6e68d"
```

### Additional context

_No response_