[Deferred] Configurable retry options

Description

Currently, Rage::Deferred automatically retries failed tasks up to 5 times using exponential backoff. However, this retry count is hardcoded and cannot be customized. Additionally, there is no way to specify which exceptions should trigger a retry.

The goal of this issue is to allow users to configure retry behavior per task:

1. Custom retry count - allow specifying the number of retry attempts (instead of the hardcoded 5).
2. Retryable exceptions - allow specifying a list of exception classes that should trigger a retry. All other exceptions would not be retried.

For example:

class SyncWithExternalService
  include Rage::Deferred::Task

  retry_on Net::OpenTimeout, Net::ReadTimeout, attempts: 10
  # or
  retry_on attempts: 3

  def perform(user_id:)
    # ...
  end
end

Tips

• Check the Deferred docs to understand how Rage::Deferred works.
• Look at the existing retry logic in the codebase to understand where the hardcoded 5 is used and how exponential backoff is applied.
• Consider what the default behavior should be when retry_on is not specified (the current behavior of 5 retries on all exceptions should be preserved for backward compatibility).
• Keep scope small: per-task config only; global config can be a future issue.
• Check the architecture doc that shows how Rage's core components interact with each other and outlines the design principles.
• Feel free to ask any questions or request help in the comments below!