Optimize FlushImagesQueue worker performance with config caching and batch processing

Problem/Motivation

The current FlushImagesQueue worker has significant performance bottlenecks that make it inefficient and slow, especially when processing large numbers of files:

• Repeated config loading: Configuration values are loaded from the database on every queue item processed via multiple $this->configFactory->get() calls
• Entity loading in loops: Image styles are loaded from the database on every queue item via $this->entityTypeManager->getStorage('image_style')->loadMultiple()
• O(n²) array operations: Using array_merge($errors, $batch_result['errors']) in loops creates new arrays and copies all existing elements on each iteration
• No batch processing: All files are processed sequentially without memory management, leading to potential timeouts and memory exhaustion
• Inefficient file operations: Multiple file_exists() calls and poor error handling can slow down processing

These issues result in poor performance for sites with many cached external images and can cause queue processing timeouts.

Steps to reproduce

1. Configure imagecache_external with a large number of external images
2. Enable cron-based flushing with a significant queue of items
3. Run the flush queue worker and observe:
   • Slow processing times (multiple database queries per item)
   • Memory usage increasing over time
   • Potential timeouts on large queues
4. Profile the queue worker to see repeated config/entity loading

Proposed resolution

Optimize the FlushImagesQueue worker with the following improvements:

1. Config caching: Load and cache configuration values once during worker construction instead of on every queue item
2. Entity caching: Pre-load image styles once and cache the directory paths to eliminate repeated database queries
3. Batch processing: Process files in configurable batches (e.g., 50 items) to manage memory usage and improve performance
4. Efficient error handling: Replace O(n²) array_merge() operations with array_push(...$errors) or immediate error logging
5. Optimized file operations: Use prepareDirectory() and reduce unnecessary file_exists() calls
6. Better exception handling: Implement granular error handling so individual file failures don't stop batch processing

These changes will significantly improve performance and scalability while maintaining backward compatibility.

Remaining tasks

• Review and test the proposed patch
• Ensure backward compatibility
• Update documentation if needed

User interface changes

None. This is a performance optimization that maintains the same external behavior.

API changes

None. The queue worker interface remains unchanged. All optimizations are internal implementation improvements.

Data model changes

None. No database schema or data structure changes are required.