Repository navigation
Conversation
|
the new failure is retained separately from carrier startup selection:
all PRs remain unmerged drafts. no deployment, live infrastructure change, release, or deployed benchmark occurred; native production admission remains disabled. |
|
disposable Linux reproduction completed on signed #484 source
the #482 CI revoked-HTTP failure remains unexplained. neither pass identifies the original response status/service, so the exact-response/control-deadline diagnostics remain in #484 and the combined #485 stack. no admission retry, timeout increase, serving change, production operation, or claim of final qualification was introduced. the integrated #485 routine run separately hit a five-second QUIC fixture connection timeout before full-duplex assertions. its focused unchanged |
|
combined source reconciliation is signed/pushed in #485 at on that combined source:
#484's isolated Linux QUIC workload reproduction also passed in 223.298s. neither it nor the combined proof explains #482's original revoked-HTTP response, the earlier manual-Compose startup failure, or the original control-outage EOF. their evidence and open qualification status remain intact. latest CI snapshot: #483 has 48 passing checks, one expected skip, and Runtime workloads still running; #485 has 43 passing checks, one expected skip, and runtime shards/workload/load still running. this is not final-matrix completion. remaining product work includes reconnect handoff (active outer closure still retires its registration), assignment/source/network-change recovery, ICE/direct upgrade, public-ingress cutover, proxy-source qualification, guest serving-tail/metrics, and remaining website service-access/source-pin reconciliation. production native admission remains disabled. nothing merged or deployed; no live infrastructure, release, or deployed benchmark operation occurred. |
What
Carry the existing control-outage/certificate diagnostics and the new exact workload-denial evidence together on the native carrier-selection stack. Stacked on #483; the workload evidence is also isolated against its failing source in #484.
Why
#470's later
61033ba5diagnostic commit was not an ancestor of #481–#483. Qualifying only those descendants therefore lost the allocation/process-lease and certificate observations that had been added for the unexplained control-outage EOF and split-startup readiness failure.#482 subsequently failed the revoked-HTTP workload proof without retaining the response status or service. Keep both evidence paths on the source being qualified, rather than treating later focused passes as explanations of either failure.
Changes
61033ba5unchanged, preserving exact non-secret allocation/process-lease deadlines and split certificate-preparation evidence.