The connections kept failing. Traffic slowed, sessions dropped, and the cluster logs screamed. The weak point wasn’t Boundary itself — it was the missing external load balancer.
HashiCorp Boundary is built to control and secure access to systems without exposing private networks. But in production, with multiple worker nodes and scaling demands, you need a robust external load balancer to keep connections steady and distribute traffic evenly. Without it, you risk uneven load, dropped sessions, and insecure network paths.
An external load balancer for Boundary serves a clear purpose: it routes client requests to healthy Boundary workers automatically, using health checks to bypass failing nodes. This reduces downtime, optimizes throughput, and simplifies DNS management. It becomes part of the security model, ensuring users hit only verified entry points.
Boundary supports TCP-based load balancing, so you can choose Nginx, HAProxy, AWS Network Load Balancer, Google Cloud Load Balancing, or other providers. Use TLS to secure connections between the balancer and Boundary workers, and configure health checks for the /health endpoint. Properly tuned, this configuration prevents stale sessions during worker restarts and upgrades.