What this tests
timeouts
Point your client at this specimen and find out — before prod does — whether it has a timeout, and whether the timeout has a timeout.
Anatomy
Akamai shipped a config bug in Edge DNS and name resolution for a decent slice of the planet quietly ceased — PlayStation Network, Steam, airlines, actual banks. The services were all fine. Perfectly healthy. Nobody could find them. It's the internet equivalent of every road sign in the country being replaced by a polite shrug overnight. To a client, DNS death doesn't look like an error at all — there's no status code, no body, no response, just a name that no longer resolves to anything. The request fails before it begins, like a letter to a street that has been renamed by decree. This specimen simulates that by holding your connection and saying absolutely nothing, which is what the network feels like from inside the failure. Your resolver will give up eventually. Your retry logic will want to. Your users already have.
Source: Akamai Edge DNS service disruption summary · 2021-07-22
Replay it
curl -i https://faultydata.com/r/FD-2026-025
// response appears hereSignature
{Status:0 Headers:map[] Body: Corrupt: Delay:<nil> Flaky:<nil> Sequence:<nil> Drip:<nil> StallAfter:0 Hang:true Reset:false Raw: Phases:[]}
Keep digging
FD-2026-006 — Total reachability loss (Meta BGP withdrawal, Oct 2021) FD-2026-043 — Datacenter literally on fire (OVH, Mar 2021) FD-2026-107 — Slow response stream (1 byte per second, forever-ish) FD-2026-003 — Elevated latency + partial 5xx (AWS us-east-1, Nov 2020) FD-2026-016 — Slow but correct response FD-2026-017 — Connection accepted, never replies