zyycnandCodex b3dd6bbe46 feat(openai): transparently retry websocket connection limit errors ...

When upstream rejects a NewChain request with websocket_connection_limit_reached
before the first deliverable event, retry once on a fresh connection and
downgrade to HTTP/SSE if the new connection is still limited or the retry
preparation fails, matching the official retry semantics. Non-NewChain
requests keep the previous behavior of delivering the error frame as-is,
and the code is persisted in upstream diagnostics. All pre-delivery
fallbacks share one helper that marks the transport decision and wait
metrics.

Co-authored-by: Codex <noreply@openai.com>

2026-08-02 03:04:06 +08:00

..

apps/gateway

test(gateway): sync config rejection fixtures to current example

2026-08-02 02:20:29 +08:00

crates

feat(openai): transparently retry websocket connection limit errors

2026-08-02 03:04:06 +08:00

migrations

feat(db): drop redundant provider_accounts_id_kind_uq index

2026-08-02 00:02:19 +08:00

Cargo.lock

feat(openai): zstd-compress responses request bodies like the official client

2026-08-02 02:27:46 +08:00

Cargo.toml

refactor(backend): finalize gateway package boundaries

2026-07-25 00:33:11 +08:00

rust-toolchain.toml

refactor: migrate backend to Rust 1.97 and edition 2024

2026-07-12 19:10:56 +08:00