Project how an Ansible inventory will grow as infrastructure scales over time.
Linear inventory growth projection: projected_hosts = current + new_hosts_per_month × months. Comparing the projected count against a scaling limit (the point where flat inventory files, single control nodes, or fixed fork settings start to strain) tells you how much runway remains before you need to invest in dynamic inventory plugins, inventory splitting, or a pull-based/event-driven automation model.
projected_hosts = current_hosts + hosts_added_per_month × projection_months
It depends on your setup, but common inflection points are: a few hundred hosts (static INI inventory becomes unwieldy), a couple thousand (single control node forks/memory starts to strain), and tens of thousands (need for inventory caching, AWX/Tower-style distributed execution, or pull-based agents).
Rarely in the long run — cloud-native environments often grow in bursts (new services, new regions) rather than steadily, so treat this linear projection as a planning heuristic, not a precise forecast, and revisit it after any major infrastructure initiative.
Move to a dynamic inventory plugin with caching, split the inventory by team/environment/region, and re-evaluate whether a control-node-per-region or pull-based model (e.g. Ansible Pull) fits better than a single push-based control node.
This model uses net growth. If you have high churn (many hosts added and removed monthly) but low net growth, the raw host count may stay flat while operational overhead (inventory refresh cost, cache invalidation) still increases — track churn separately if it's high.