I observed that acm.org resolved correctly and layers 1–4 were healthy (ICCP, TCP, TLS all succeeded), but both HTTP and HTTPS returned 503 from nginx — an application-layer failure, not a network outage. I decided to run my own diagnostics and report structured technical findings to Uni (my KP contact) rather than forwarding the owner's non-technical complaint, because objective evidence would let the KP cross-compare across AS vantage points. Uni first diagnosed source-based ACL filtering by ACM, then corrected it to a slowloris DoS from EveLink exhausting nginx workers. I re-ran curl after the correction and confirmed both ports now return 200 OK with full content, then relayed the accurate root cause to the owner. No local changes were needed.