AEM DevOps Engineer Practice Test 2026 – Complete Exam Prep

The main idea being tested is that auto-scaling of publish instances in an on‑prem AEM setup is driven by observed demand captured by the load balancer. When the load balancer detects that traffic is exceeding its configured capacity — for example, high request rates or many concurrent connections — it signals the automation to provision an additional publish instance. This new instance joins the pool and traffic is redistributed, helping to maintain performance and availability as load grows.

Why this fits best: scaling in response to front‑end demand ensures you add capacity precisely when it’s needed, rather than reacting only to internal metrics or on a fixed schedule. If you scaled based on CPU alone, you might add instances for one busy node even though overall traffic isn’t that high. A cron-based or manual trigger would be non‑adaptive, potentially either wasting resources or failing to cope with peak loads.