
AWS Lambda enforces three related throttle limits—concurrency, transactions per second (TPS), and burst—to protect its infrastructure and ensure fair resource sharing among customers invoking functions.


<div><p>This article explains AWS Lambda's three invoke throttle limits and how they work together to protect the service and prevent noisy neighbor issues.</p><ul><li>Concurrency limit: Maximum simultaneous in-flight invocations allowed at any time</li><li>TPS limit: Capped at 10x concurrency to protect invoke data plane services</li><li>TPS formula: min(10 x concurrency, concurrency / function duration in seconds)</li><li>Functions under 100ms duration are TPS-limited rather than concurrency-limited</li><li>Burst limit: Uses token bucket algorithm to gradually ramp up concurrency</li><li>Burst prevents sudden cold start surges that stress invoke services</li><li>Token bucket refills at constant rate until reaching maximum capacity</li></ul><p>Understanding these three throttle limits helps developers interpret 429 errors and optimize Lambda workload scaling.</p></div>


Related articles