- Add express-rate-limit for OAuth initiation (10 req/min) and callback (30 req/min)
- Skip rate limiting in development/test environments
- Update tests with rate limiter mock

Addresses PR #1141 Follow-up #3: OAuth rate limiting

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

- Add express-rate-limit for OAuth initiation (10 req/min) and callback (30 req/min)
- Skip rate limiting in development/test environments
- Update tests with rate limiter mock

Addresses PR #1141 Follow-up #3: OAuth rate limiting

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

- Add express-rate-limit for OAuth initiation (10 req/min) and callback (30 req/min)
- Skip rate limiting in development/test environments
- Update tests with rate limiter mock

Addresses PR #1141 Follow-up #3: OAuth rate limiting

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Test #3 from constructive-planning#911: full metering pipeline tests.

Metering E2E tests (15 tests):
- Config resolution: getLlmBillingConfig reads metaschema tables
- Billing functions: check_billing_quota + record_usage SQL stubs
- Inference log: logInferenceUsage writes correct fields
- meteredEmbed with mock embedder: quota check, record_usage, quota exceeded, request_id propagation
- meteredEmbed with real Ollama: live inference + billing pipeline (skipped when Ollama unavailable)

CI Tier 4 (ollama-tests):
- PostgreSQL + Ollama service containers
- Pulls nomic-embed-text model before test run
- Runs graphile-llm full test suite including Ollama E2E