herjarsa c347c34af0 fix(admission): queue heavyweight chat requests before 503 busy ...

Agent clients (OpenCode, Claude Code, Cursor) fan out heavy sub-requests
that land on the admission gate together. With the single heavyweight
slot, concurrent heavy requests were rejected immediately with a
retryable 503; clients burn their retry budget in seconds and the agent
dies mid-task.

Heavy requests now wait up to OMNIROUTE_CHAT_ADMISSION_QUEUE_MS (default
5000ms) for a slot before the 503, served FIFO; 0 restores the legacy
immediate-reject behaviour. Applied to both the byte-based path
(admitChatRequest) and the structure-based path (admitChatStructure, now
async).

2026-08-09 01:33:24 -03:00

..

API_REFERENCE.md

feat(guardrails): modality bridge core — vision mode/task-aware/cache/input_image + modalityBridge settings (#9759)

2026-08-08 01:59:11 -03:00

CLI-TOOLS.md

Release v3.8.49 (#7076)

2026-07-29 15:18:55 -03:00

ENVIRONMENT.md

fix(admission): queue heavyweight chat requests before 503 busy

2026-08-09 01:33:24 -03:00

FEATURE_FLAGS.md

Release v3.8.49 (#7076)

2026-07-29 15:18:55 -03:00

FREE_PROXIES_API.md

Release v3.8.47 (#6569)

2026-07-13 09:12:40 -03:00

FREE_TIERS.md

Release v3.8.49 (#7076)

2026-07-29 15:18:55 -03:00

meta.json

Release v3.8.47 (#6569)

2026-07-13 09:12:40 -03:00

PROVIDER_PLUGIN_MANIFEST.md

Release v3.8.49 (#7076)

2026-07-29 15:18:55 -03:00

PROVIDER_REFERENCE.md

fix: update Baichuan website URL to baichuan-ai.com (#9312)

2026-08-05 21:46:11 -03:00

RELAY_BACKEND_STRATEGY.md

Release v3.8.44 (#5925)

2026-07-04 13:00:30 -03:00

RELAY_TROUBLESHOOTING.md

Release v3.8.47 (#6569)

2026-07-13 09:12:40 -03:00