Hi HN, I’m building Fairvisor — a programmable rate limiting and API governance layer. Most rate limiters today feel operationally incomplete: Redis counters glued onto gateways hardcoded configs weak visibility into why requests were blocked little connection between traffic policy and business logic Fairvisor treats rate limiting as a governance problem instead of just request counting. Current focus: centralized policy management edge enforcement tenant/country-aware rules explainable decisions observability integrations programmable traffic policies I’m also experimenting with a lock-free/event-driven dataplane instead of the classic “everything goes through Redis” architecture. Some example policies: AI agents may consume at most N% of shared capacity expensive endpoints automatically tighten limits under load different limits by geography, tenant, or plan soft throttling instead of binary blocking Mainly interested in feedback from people operating public APIs or AI infrastructure. Especially curious: Do you think API governance becomes its own infrastructure category, or are existing gateway tools already enough?