feat(openai-router): use per-model retry config from WorkerRegistry

[CatherineSue]

Description

Part of the per-worker resilience refactor series: #799 → #803 → #821 → #836 → #875 → #881 / #933 → this PR.

Problem

OpenAI router uses a single global retry_config for all requests, ignoring per-model retry config set by workers.

Solution

Resolve per-model retry config from WorkerRegistry in route_chat() before passing it to ChatRouterContext, falling back to the router-level default.

Independent of #881 (HTTP router) and #933 (gRPC router).

Changes

• route_chat() in OpenAI router resolves retry config from WorkerRegistry per model ID

Test Plan

• cargo test -p smg --lib — all 450 tests pass
• Pre-commit hooks pass (rustfmt, clippy, codespell, DCO)

Checklist

• cargo +nightly fmt passes
• cargo clippy --all-targets --all-features -- -D warnings passes
• (Optional) Documentation updated
• (Optional) Please join us on Slack #sig-smg to discuss, review, and merge PRs

Summary by CodeRabbit

• Bug Fixes
  • Enhanced retry configuration handling by enabling per-model resolution. The system now applies model-specific retry configurations when available, with automatic fallback to default settings, improving overall reliability and resilience.

Summary by CodeRabbit

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro

Run ID: d82c5b85-c4d7-4604-9850-f59767745676

📥 Commits

Reviewing files that changed from the base of the PR and between f04d3ee and 4ef1e59.

📒 Files selected for processing (1)

• model_gateway/src/routers/openai/router.rs

---

📝 Walkthrough

Walkthrough

The OpenAI router's chat routing method now resolves per-model retry configuration from the worker registry, with fallback to the router-level default configuration if unavailable. The chat routing context is updated to receive this selected configuration instead of the router's global retry config.

Changes

Cohort / File(s)	Summary
OpenAI Router Retry Configuration model_gateway/src/routers/openai/router.rs	Updated route_chat to resolve per-model retry config via worker_registry.get_retry_config(model_id) with fallback to router-level default; passes selected config to ChatRouterContext.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

Possibly related PRs

• feat(grpc-router): use per-model retry config from WorkerRegistry #933 – Implements the same per-model retry configuration pattern for the gRPC router, providing a parallel implementation for consistency across routing layers.
• feat(core): wire per-worker resilience and HTTP client into BasicWorker #803 – Introduces per-worker resilience and retry configuration wiring that enables the worker registry lookups used in this change.
• refactor(gateway): split OpenAI router.rs into chat and health modules #726 – Introduced the ChatRouterContext structure containing the retry_config field that this change modifies to accept per-model configurations.

Suggested labels

model-gateway, openai

Suggested reviewers

• key4ng
• slin1237

Poem

🐰 Per model, the retry config now flows,
From worker registry where wisdom grows,
With fallback grace when lookup shows no trace,
The router routes with smarter pace! ✨

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and specifically describes the main change: the OpenAI router now uses per-model retry configuration from WorkerRegistry instead of a single global config.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/openai-router-retry-from-registry

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[gemini-code-assist]

Code Review

This pull request implements per-model retry configurations in the OpenAI router, allowing for model-specific retry logic with a default fallback. A review comment suggests refactoring the configuration retrieval to return references to avoid unnecessary clones and improve efficiency.