Advanced load balancing for Azure Application Gateway
is now available in public preview, giving you greater control over how traffic is
distributed across backends supporting different workloads, including AI
applications and web traffic. New server-level algorithms and backend pool
controls let you tailor traffic distribution to your application’s capacity and
routing requirements.
At the backend pool level, priority controls let
you designate preferred pools, while weights let you configure the
relative traffic share across eligible pools. For AI applications, you can
prioritize pools hosting preferred large language model (LLM) deployments. For
web applications, you can adjust traffic distribution across pools to reflect
differences in backend capacity and deployment needs. Within each backend pool,
Least Connections directs traffic to servers with fewer active
connections, while Weighted Round Robin distributes traffic according to
configured server weights. Together, these capabilities provide flexible,
policy-driven traffic management for AI and web workloads.
Learn more.