Problem

We currently have no way to know if a deployment can reach our usage reporting system until they actually send usage data. This makes it difficult to distinguish between customers blocked by firewalls versus customers with low/no usage.

Solution

Add a heartbeat mechanism where deployments send a periodic ping to indicate they can reach the usage reporting system.

• Send lightweight payload (e.g., "I can reach you") on a regular interval
• Use this data to identify firewall issues during POC and post-deployment
• Helps CS team proactively identify connectivity problems

Related Issues

• https://github.com/coder/nexus/issues/256
• Change default AI workspace limit to 1000 per deployment #1106