how do you design for graceful degradation when services fail, dont you?

hey folks, ive been thinkin a lot abt what happens when one of our upstream services decides to take an unscheduled nap. beyond just standard retries, how do you architect applications to degrade gracefully? are you thinkin abt circuit breakers, fallback data, or maybe even feature flags to disable non-critical paths? curious to hear what patterns and tools youve found most effective in keepin the user experience decent, even when parts of the system are down