QStash overhauled flow control to become a runtime-managed feature rather than a publish-time configuration. The rewrite addressed a core usability issue where cancelling rate-limited messages left slots blocked, making the feature practically unusable—now messages sit in an ordered wait list instead of being pre-scheduled to future slots, so cancellations free capacity immediately for new publishes. Alongside this, the team added pause/resume operations and an API to pin rate and parallelism values outside the publish path, plus expanded the flow control metrics endpoint to surface waitlist depth, active task counts, and effective limits so users can diagnose why messages queue. Two other threads emerged: sensitive data protection via field-level redaction in publish calls (body and header whitelisting), and regional expansion with US and EU endpoints to reduce latency.
QStash expanded flow control capabilities and added data protection features. The team shipped pause/resume controls for flow-control keys, a new API to manually pin and adjust rate and parallelism settings without republishing messages, and enhanced the get flow-control endpoint with detailed metrics showing parallelism count, rate usage, and waitlist size to help users monitor limits. Sensitive data redaction also shipped, letting developers mark message bodies and specific headers to be redacted from the dashboard and API responses. Regional expansion to US and EU endpoints reduced latency for message ingestion.