Consolidate request diagnostics into typed request_observation_json with
migration 0025, preserving historical facts and query constraints.
Use upstream response timestamps for output throughput, align TTFT event
boundaries, and surface official timing metrics in the latency popover.
Co-authored-by: Codex <noreply@openai.com>
Give Guardian requests a separate per-account lease pool and wait queue while preserving ordinary concurrency and shared request intervals. Use the ordinary pool when the Guardian limit is zero, and align scheduling, diagnostics, settings copy, and documentation.
Co-authored-by: Codex <noreply@openai.com>
Check the recorded native continuation scope before selecting an account
or committing a session binding. Reuse the same scope validation after
middleware while preserving native connection checks and migration policy.
Cover ReplayOwner and ReplayAny with persisted provider metadata or a
connection-local pin, with and without an existing session binding.
Rejected replays must not acquire leases, write bindings or aliases, send
upstream requests, or redirect subsequent child requests.
Refs #364
Co-authored-by: Codex <noreply@openai.com>
Allow requests with a reliable session identity to claim a missing account
binding, including child threads and independent background tasks. Remove
the missing-binding wait queue while preserving existing migration rights
and bound-account waiting behavior.
Cover concurrent root/child claims, memory prewarm and turn reuse,
middleware identity changes, cancellation, and bound-account timeouts.
Update the documented strict-affinity contract.
Co-authored-by: Codex <noreply@openai.com>
Separate execution, provider ports, administration, and plugin lifecycle responsibilities.
Fix shutdown ordering, account lock reclamation, multipart cancellation, budget windows, and WebSocket close handling.
Align xAI catalog, reasoning, tools, replay, quota, and cost handling with current Grok contracts and remove legacy fallbacks.
Update API and architecture documentation with regression coverage. Validated backend suites, strict workspace Clippy, isolated PostgreSQL/Redis integration, and real Codex/Grok client flows.
Co-authored-by: Codex <noreply@openai.com>
Move observability query policies into Admin and reuse account state reads.
Bound Redis reads to measured batches and share protocol and transport helpers.
Close backup process and staging cleanup gaps, and remove unused entry points,
redundant models, dependencies, and obsolete tests.
Validated 3,526 workspace tests, then all 352 Store tests after batch tuning,
plus strict workspace Clippy and formatting checks. Local release benchmarks
cover concurrent account reads and controlled transport delays.
Co-authored-by: Codex <noreply@openai.com>
Count non-preamble Responses events for first-token timing while preserving
semantic output boundaries and request-level timing origins. Calculate
output throughput over the complete request duration in usage rows and
aggregate queries.
Co-authored-by: Codex <noreply@openai.com>
Route relaxed requests through the configured scheduler and keep the session primary during preferred-mode spillover. Default new settings to strict while preserving saved modes.
Co-authored-by: Codex <noreply@openai.com>
Use read-only status checks for initial authentication and remove activity-driven renewal. Keep failed and newly queued requests waiting for the current refresh before replaying with the updated cookie.
Co-authored-by: Codex <noreply@openai.com>
Merge error details and account affinity settings into migration 0023. Preserve all migrations shipped in v3.20.0 and validate upgrades from its 0022 baseline.
Co-authored-by: Codex <noreply@openai.com>
Reuse domain models for observability, account statistics, execution facts,
and group validation. Compose shared settings values while preserving
HTTP and SDK contracts and explicit storage and runtime boundaries.
Verify serialization, concurrency, persistence, and snapshot behavior with
3528 backend tests, strict Clippy, formatting, and architecture checks.
Co-authored-by: Codex <noreply@openai.com>
Add relaxed and strict affinity modes, configurable binding retention, and per-request account rotation limits. Freeze policies for each request and preserve continuation ownership and plugin selection behavior.
Expose the controls in a dedicated affinity settings card and cover scheduling, persistence, renewal, and retry boundaries.
Co-authored-by: Codex <noreply@openai.com>
Preserve native failure causes through store, provider, plugin, admin, and
worker boundaries, including secondary cleanup failures. Route operational
failures through the shared diagnostics sink with protected error details.
Use bounded nonblocking file logging with batched writes and reserved
warning/error capacity. Reserve observation capacity for failures and
request lifecycle records, and rate-limit overload warnings.
Validate error propagation and blocked-output HTTP, admin, SSE, and WebSocket
behavior; keep the local migration plan out of version control.
Co-authored-by: Codex <noreply@openai.com>
Retain preparation and account-selection diagnostics, preserve plugin fault
details, and record failures before attempt startup. Keep failure
classification when trace summaries are truncated.
Carry native error sources through Core, infrastructure availability errors,
OpenAI credential selection, provider inference, and plugin upstream adapters.
Preserve original upstream codes independently of failure classification and
project bounded cause chains and upstream bodies into protected error details.
Rename persisted and admin API error detail fields, keep raw details out of
ordinary logs and diagnostic exports, and account for retained error data in
the observation queue memory budget. Add regression coverage for source
retention, safe output, and migration upgrades; remove account-dependent live
tests. Document the implemented diagnostic boundaries.
Refs #357
Co-authored-by: Codex <noreply@openai.com>
Truncate overflowing plan names and show the full display name on hover while preserving the existing subscription column width.
Co-authored-by: Codex <noreply@openai.com>
Record semantic output from the first stream chunk and reuse Core timing for xAI. Add regression and opt-in live transport coverage.
Co-authored-by: Codex <noreply@openai.com>
Renew active administrator sessions with atomic Redis expiry updates,
synchronized cookies, and idle and absolute lifetime limits. Preserve
revocation after logout and password changes.
Coordinate session recovery, bounded retries, error handling, and raw
responses through request plugins. Gate initial navigation in the router
and organize frontend API modules by their views.
Co-authored-by: Codex <noreply@openai.com>
Merge refresh margin and group Fast mode changes into migration 0022 before v3.20.0. Preserve the v3.19.0 migration history and verify upgrades from migration 0021.
Co-authored-by: Codex <noreply@openai.com>
Share the current account across root and descendant requests, including
Search, Images turn aliases, and Live creation. Keep root recovery and
explicit plugin scheduling intact while descendant requests wait for the
current owner and follow root migrations.
Use versioned admission before sending so concurrent claims and late
responses cannot overwrite newer bindings. Cover queue lifecycle,
continuation recovery, endpoint routing, and Redis concurrency behavior.
Co-authored-by: Codex <noreply@openai.com>
Freeze the selected client profile for Live sideband, reconnect and hangup,
while reloading current account credentials and outbound proxy settings.
Share identity-header protection, clear turn state across wire entry points,
and include Guardian and residency headers in WebSocket connection matching.
Normalize the existing Cargo.lock http reference without changing versions.
Validated 1389 tests, workspace strict Clippy, Rustfmt and architecture checks.
Co-authored-by: Codex <noreply@openai.com>
Stagger credential refresh and quota observations, bound each request to three account rotations, and expose terminal and version lag configuration.
Retain main's bounded refresh 401 recovery policy. Version lag accepts integers from 1 to 10 or an empty value.
Co-authored-by: Codex <noreply@openai.com>
Add scoped call bootstrap, pinned-account sideband relay and hangup. Preserve model metadata, use standard WebSocket handshakes, and release session claims safely on cancellation.
Validated with regression tests, CI and a real OAuth/WebRTC call.
Co-authored-by: Codex <noreply@openai.com>
Add Chinese responsibility summaries to every backend Rust source, test,
and build script. Remove sentence-ending periods from comments while
preserving meaningful dots in URLs, versions, identifiers, and examples.
Keep generated SDK documentation in sync with its contract generator,
group middleware module declarations and exports, and record the comment
conventions in the development skill.
Validated with Rustfmt, strict Clippy, backend tests, documentation tests,
and skill validation. Database-backed integration cases were skipped
because dedicated PostgreSQL and Redis test services were not configured.
Co-authored-by: Codex <noreply@openai.com>
Move shared components and helpers to their actual owners, standardize
component directory names, and use concise names for common utilities.
Remove redundant wrappers, copied API models, and duplicate state.
Separate API key editing from key usage, isolate settings drafts, reuse
paged queries, and tie debounced queries and requests to their scopes.
Validated with frontend lint, type checking and production build, local
import checks, and 15 temporary behavioral regression groups.
Co-authored-by: Codex <noreply@openai.com>
Keep the glibc tuning value in config.yaml's app-runtime bridge and let
Compose inject it before the application starts. Leave tuning disabled
by default and document the 128 KiB preset, recreation, and upgrade steps.
Refs #338, #345
Co-authored-by: Codex <noreply@openai.com>
Use event-listener registrations that unregister when cancellation futures
are dropped, including waits attached to ancestor tokens. Recheck the
cancelled flag after registration to preserve wakeups during races.
Add allocation regressions for select churn and child cancellation, plus
broadcast and concurrent cancellation coverage.
Refs #338, #345
Co-authored-by: Codex <noreply@openai.com>
Allow long model responses to finish without a fixed 600-second deadline,
while preserving explicit plugin deadlines and upstream idle timeouts.
Renew request recovery, client admission and provider scheduling leases
for the session lifetime, and release their renewal tasks with the owners.
Keep observation renewal best-effort and cancel on lost concurrency leases.
Fixes#342
Co-authored-by: Codex <noreply@openai.com>
Configure Guardian reservations through persisted runtime settings and frozen request policies. Combine concurrency controls and queues in the admin UI, with the scheduling strategy shown first.
Prioritize Guardian account waiters, recheck displaced queue heads before lease acquisition, and report request-specific available capacity. Preserve account, key and wait-budget limits.
Co-authored-by: Yongzhao Chen <yongzhao.derek@gmail.com>
Co-authored-by: Zyy <zyycn@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Codex <noreply@openai.com>
ZyyandCodex
KumikoLoveandCodex
2han9wen71anandCodex
dependabot[bot]
WeiSiTaPanandCodex
