Retry loop ignores Retry-After header

Your service calls a vendor API that rate-limits with HTTP 429 + a Retry-After header. At 12:00 you cross the limit during a busy period and start getting 429s, expected, should self-resolve as the window resets. Instead the 429 rate stays pegged near 100% for 20+ minutes even as organic demand falls. Dashboards: your outbound request rate to the vendor stays HIGH and flat throughout, and your client retries every 429 immediately (tight loop, no delay, ignores Retry-After). How do you triage and mitigate?