Knowledge base · TOPIC-19

Changing egress IP: failure or normal multi-egress?

Short answer: not necessarily a failure. A logical service or network path may use multiple exits. Exact IP, address family, country, provider network, and application continuity are different evidence dimensions.

One observation can show that an endpoint changed; it cannot establish why. Repeated samples may reveal a small stable pool or a dynamic set, but neither is automatically good or bad. Application success and the expected route policy must be evaluated separately. Never treat “same country” as permission to trust every network, and never publish a user’s exact public IP as a diagnostic badge.

Collect only consented, time-bounded samples. Separate endpoint rotation from selected-route changes and from protected traffic going direct. Preserve ASN or organization only when justified and redact exact addresses from public exports.

Open Probe V0.1 does not rank exits, attribute a provider outage, or issue a commercial reliability score. It can record local evidence under its privacy contract. Use Open Probe for explicit local measurement and the Scanner for a limited browser snapshot. Related: IPv4 healthy while IPv6 is degraded.