Optimize fork sync schedule and investigate impact on rate limit

Relevant Slack thread in #localization-engineering: https://gitlab.slack.com/archives/C08JRUR75M4/p1760454738437549

Current state

We are experiencing intermittent "Failed to fetch fork details. Try again later." errors on our localization forks. This appears related to Cloudflare rate limits on the GraphQL API for the getForkDetails query, tracked in https://gitlab.com/gitlab-org/gitlab/-/issues/566800+

• 12 forks total across Test and Prod environments
• Hourly sync schedule for all forks
• Error messages appear intermittently but not critically blocking continuous translation work across Translation MRs

To-do

• Understand if Localization team's fork usage (use case is, multiple large forks with frequent syncing) is creating unintended load
• Understand what is the actual compute/resource cost of our current hourly sync pattern
• Optimize prod fork sync schedule (reduce sync frequency, stagger sync times, reduce or remove the syncing for test forks
• Document the sync configuration on https://handbook.gitlab.com/handbook/marketing/localization/engineering/#syncing-product-documentation-forks-with-upstream-projects