Congo Proxies

Begin with the smallest harmless request that proves the configuration. Preserve permission evidence and the effective scope beside the run, and stop when a quota, denial, challenge, access-control message, or owner instruction says to stop. The COD audit trail should pair each request ID with allowed source, client configuration, network and protocol, session behavior, requested country, observed exit, response classification, time, and retry relation. Preserve stop signals without circumvention. Use only an endpoint permitted for the Congo test, keep missing measurements explicit, and never rotate around a refusal or other mandatory stop.

Separate local resolver errors, proxy-side lookup failures, connection timeouts, and destination responses. A country-targeted exit does not by itself establish the resolver path, and a successful lookup does not prove the requested exit country. The Africa record immediately on either side of Congo is represented by Malawi at 11,354 and Cape Verde at 10,764. This is a historical-count comparison. It is not a recommendation to substitute countries or a measurement of current service. Keep COD comparator rows labelled and diagnose the requested Congo route from its own evidence rather than a neighbouring count. Record why each comparator belongs in the authorized review before using it.

Approve the requested country only from its own verified pilot. A successful regional control can explain the experiment, but it cannot replace missing evidence for the route the buyer intends to use. Congo buyers should reject any COD result that lacks exit verification, protocol compatibility, session evidence, a stable acceptance rule, failed-attempt accounting, or a bounded retry policy. Approve only the configuration that passed the recorded pilot. Reject any Congo conclusion built from an unverified exit, hidden failures, or a changed rule, and repeat COD verification after a material configuration change.