Extract Gitlab::Database::Batch::Optimizer

Ref: #561698 (closed)

What does this MR do and why?

This MR refactors the batch size optimization logic for batched background migrations into a pure service class pattern.

Previously, the optimization logic was tightly coupled within a class. Now it's split into:

• Gitlab::Database::Batch::Optimizer - a pure service class that calculates optimized batch sizes without side effects
• BatchedMigration#optimize! - orchestrates the optimization and handles persistence

This separation allows:

• Better testability with pure functions
• Clearer separation of concerns (calculation vs persistence)
• Flexibility in how optimization is applied

The optimizer adjusts batch sizes based on time efficiency to keep migrations running at optimal speed (90-95% efficiency), automatically increasing batch size when migrations are too fast and decreasing when they're too slow.

migration = BatchedMigration.find(id)
migration.optimize! # Calculates and updates batch size if needed