The proxy failed at 3 a.m., and the entire private subnet went dark. Five minutes later, half your internal tools were unreachable. That’s when you realize the flaw wasn’t the proxy—it was the architecture.
A VPC private subnet proxy can be the gatekeeper between your internal-only systems and the outside world, but without domain-based resource separation, you’re piling too much trust into a single point. Split your domains, isolate resources, and let your proxy enforce clear boundaries. When designed right, each domain maps to its own isolated backend, each with tailored permissions, security groups, and routes.
The power of domain-based separation lies in containment. A compromise in one area doesn’t cascade into another. With a proxy in a private subnet, DNS rules can route traffic to distinct destinations based on domain names, removing the need for blanket network access. This reduces lateral movement risks, simplifies auditing, and tightens your access policies without slowing anything down.
For deployment, start with a VPC configured for at least two tiers of subnets—public for load balancers, private for proxies, backends, and databases. Place your proxy inside the private subnet with no direct inbound access from the internet. Add a layer of domain-based routing in the proxy configuration. Assign each domain to its own target group or upstream resource. Store these mappings in code, version-controlled, and load them dynamically for fast changes without redeploying infrastructure.