What problem does it solve? Okta Workflows event hooks, streaming helper flows, and API Endpoint flows have undocumented or surprising hard limits and behavioral quirks that cause flows to silently fail, drop connections, or lose state across pages. This Skill consolidates those constraints and their workarounds so you can design hook and handler flow architectures that survive production limits. ## Core Features & Use Cases - Streaming flow constraints: Explains why streaming helper flows cannot be stopped or deactivated mid-run, the 10,000-record standard search limit versus the 1 million-record streaming maximum, and the rigid requirement that helper flow inputs be named exactly Record and State (both Object). - Pagination workarounds: Documents why State is not a mutable cross-page accumulator and how to use the Connector Builder Paginate card's object/break/path 'Do while' pattern with its 5,000-iteration cap, plus externalizing state to Tables rows. - Event hook and API Endpoint limits: Covers the 3-second event hook timeout with single retry, 400,000 events per 24 hours, 25 active hooks, no ordering guarantee, and the 60s sync / 120s async connection drops with API Connector Close and Call Flow Async workarounds. - Use Case: When building a streaming flow to process 500,000 user records, use this Skill to plan for the ~13-hour runtime, structure Record/State inputs correctly, and persist cross-page state in a Tables row keyed by run ID. ## Quick Start Ask how to design an Okta Workflows streaming helper flow that paginates an API and persists state across pages without hitting platform limits.
