dudukee b5bd63b585 fix(openai): use a catalog-supported warmup reasoning effort (#425) ...

Scheduled account warmup sent reasoning.effort=none, but the Codex model
catalog lists low as the lowest supported level for every model an account
can see. Upstream answers HTTP 200 with an SSE error event (unsupported_value
on reasoning.effort), which the warmup terminal check records as
response_failed, so the cycle ends with warmed_up=0 and never starts the 5h
window.

Send low instead and cover the upstream payload with a regression test.

2026-10-14 11:42:48 +08:00

..

gateway-admin

fix(auth): separate access and refresh session credentials (#426)

2026-10-14 11:42:33 +08:00

gateway-api

fix(auth): separate access and refresh session credentials (#426)

2026-10-14 11:42:33 +08:00

gateway-core

fix(openai): forward response preludes without buffering

2026-10-14 01:51:53 +08:00

gateway-host

fix: preserve terminal failure and cleanup causes (#417)

2026-10-13 11:31:10 +08:00

gateway-plugin

feat(settings): remove account request interval limits

2026-10-13 21:50:53 +08:00

gateway-protocol

perf(store): scale request analytics and isolate request history (#389)

2026-10-11 20:57:36 +08:00

gateway-store

fix(auth): separate access and refresh session credentials (#426)

2026-10-14 11:42:33 +08:00

providers

fix(openai): use a catalog-supported warmup reasoning effort (#425)

2026-10-14 11:42:48 +08:00