What problem does it solve? Event-driven functions can overwhelm external APIs, duplicate work from rapid event bursts, and create unfair resource contention across tenants. This Skill provides prescriptive guidance for applying Inngest flow control primitives to manage load, prevent duplicates, and prioritize important runs. ## Core Features & Use Cases - Concurrency & Throttling: Limit simultaneous step execution with per-key or account-level scopes, and spread function starts over time to respect API rate limits like OpenAI or HubSpot quotas. - Debounce, Singleton & Rate Limiting: Collapse rapid event bursts into a single run, ensure only one instance runs per key, and discard excessive duplicate events. - Priority & Batching: Give VIP or critical runs queue priority based on event data, and process events in groups for bulk efficiency. - Use Case: A multi-tenant SaaS hitting OpenAI 429 errors can combine account-level concurrency limits, per-user fairness keys, and priority expressions so all customers share API quota fairly while paid plans get faster processing. ## Quick Start Ask the AI to add Inngest flow control to a function that keeps hitting OpenAI rate limits, using throttling and per-user concurrency keys.