Rate Limiting vs Load Shedding vs Circuit Breaking

A checkout service calls a third-party tax-calculation API. During a regional incident:

• The tax API starts timing out on ~80% of calls.
• Simultaneously, checkout's own CPU climbs to 95% because retries are piling up.
• Meanwhile, one particular customer's storefront is sending 10× its normal traffic due to a bug in their integration.

1. For each of the three problems, name the mechanism (rate limiting, load shedding, or circuit breaking) that addresses it, and explain why the other two don't.
2. Explain how, left unaddressed, the second problem (retries piling up) is actually made worse by the first problem, and how a well-designed system breaks that loop.
3. Could all three mechanisms be active in the same request path at once? Give an example ordering.