Add documentation for pipeline processing retry behavior and self-healing

This change improves the documentation for GitLab's CI/CD pipeline processing system. The main updates include:

1. Removed outdated content: Deleted a reference to an external diagram link and condensed a lengthy paragraph about pipeline processing into a shorter, cleaner version.
2. Added detailed explanations: Created new sections that explain how pipeline processing actually works behind the scenes, including when it runs automatically and how it manages job status changes.
3. Clarified job lifecycle: Better explained how jobs move through different states (like "created" and "pending") and what triggers these transitions.
4. Added reliability information: Included a new section about how the system handles errors and recovers from temporary problems using retry mechanisms with increasing delays.

The changes make the documentation more comprehensive and easier to understand by breaking down complex technical processes into clearer explanations while removing unnecessary details. The overall goal is to help developers better understand how GitLab's pipeline system processes and manages CI/CD jobs from creation to completion.