The Azure Functions Flex Consumption plan adds VNet support and per-instance concurrency control

Flex Consumption became generally available in 2024. It keeps scale-to-zero and per-second billing while adding virtual network integration, a configurable instance memory size, and always-ready instances to reduce cold starts, capabilities that previously required the Premium plan.

It also exposes an explicit per-instance concurrency setting, so you decide how many concurrent executions one instance handles instead of relying on the runtime's own heuristics. That matters for memory-bound workloads where the default concurrency causes out-of-memory restarts.

Before moving an app, check three things: regional availability, which is narrower than classic Consumption; the deployment model, which uses a package in a blob container configured per app rather than the classic content share; and settings that no longer apply, including the content share connection string and the dynamic scale-out limit. Verify against the plan's own documentation for your region rather than assuming parity with Consumption.