Change default AI workspace limit to 1000 per deployment

[bpmct]

Problem

Currently, the default behavior for AI workspace limits uses a per-user calculation (e.g., 800 included per user, 1000 limit per user). This creates scaling concerns where the total free usage pool can grow to unsustainable levels as the user count increases across deployments.

Solution

Change the default AI workspace limit to a flat 1000 per deployment (instead of per-user multiplication) for deployments without an explicit license limit set.

Related issues

• https://github.com/coder/nexus/issues/256
• Add heartbeat for usage reporting for Tallyman/Metronome #1107