https://nalar.dev/tags/sre/ Recent content in SRE on Nalar Hugo en-us Wed, 02 Sep 2026 00:00:00 +0700 https://nalar.dev/load-shedding-and-overload-protection-for-cloud-services/ Wed, 02 Sep 2026 00:00:00 +0700 https://nalar.dev/load-shedding-and-overload-protection-for-cloud-services/ <p>Autoscaling is useful, but it is not instantaneous. Traffic can rise faster than new instances start, a dependency can slow down, or a retry storm can multiply work. When demand exceeds safe capacity, accepting every request can make the entire service slower until almost nothing completes.</p> <p>Load shedding is the deliberate rejection or degradation of work to keep the system inside a recoverable operating range.</p> <h2 id="overload-is-often-a-queueing-problem">Overload is often a queueing problem</h2> <p>Imagine a service that safely handles 200 concurrent requests. A downstream dependency slows from 50 ms to 2 seconds.</p>