fix(sse): sticky head + store drain for opencode rotation (opt-in)

[maxmad64bis]

⚠️ base-red inherited: #13866

Summary

Concurrent requests on the opencode rotation start from the same member and replay the same failing cascade. This PR reuses the last healthy member and skips recently refused members while the opt-in skip flag is on; with the flag off the order is exactly the plain rotation. The saturated fallback serves without recording sticky, so sticky keeps the previous healthy serve.

Related Issues

• Related to fix(quota): dedup concurrent getSaturation misses with singleflight #12787

Validation

• Change type: provider
• Focused tests and category gates from the golden path
• npm run lint — local eslint on the touched files is clean; the full run is red on the base (files outside this PR)
• Reconciled with the current active release base; focused checks rerun afterward
• Production-code changes include a new or updated automated test in this PR

Tests Added Or Updated

• tests/unit/account-rotation-sticky.test.ts (new, 11 cases: dispersion, sticky, store drain, flag-off terminal, saturation, eviction, success interplay, custom predicate, 15-member wave).

Coverage Notes

• open-sse/executors/accountRotation.ts is covered by the new test plus account-rotation.test.ts and account-rotation-lot-c.test.ts, green without modification (57 tests total with the refusal-memory and empty-rejection suites).

Reviewer Notes

• Local gates on PR files are green: unit oracle 11/11, cycles OK (438 files), eslint exit 0, typecheck-core exit 0, file-size OK, complexity new-code 0 (the one hit, isEmptyUpstreamRejection: 20, is pre-existing on the base).

• CI reds are all inherited (files outside this PR, unchanged since merge-base 7a921299, same failures on fix(sse): park request with heartbeat after repeated transient 429s, replay one capped leg #13924/feat(sse): per-egress pacing + fleet-wide backoff for opencode rotation (opt-in) #14180, open base-red on release/v3.8.51): API Typecheck (rerankProviderNodes, antigravity TS2677), skills-sync (omni-version-manager generated), ESLint (cliRuntime, arena-elo, translator-openai-to-gemini), env-doc-sync, unit shards.

• CI reds are inherited from the red base (🔴 Release branch not green: release/v3.8.51 #13866), not from this diff: the third-party PR fix(sse): treat antigravity empty completions with a normal stop as valid 200s (#14160) #14243 on the same base release/v3.8.51 fails the same 9 jobs (API Route Typecheck, Docs Gates, Fast Quality Gates, Merge integrity, ESLint, Unit fast-path 1-4/4); every file cited by the failing gates is outside this diff. Non-blocking for this PR.

[diegosouzapw]

Thanks @maxmad64bis — merging via the release merge-train. Validated in local merge-train (mt-train10c) on the devbox @ train tip 4d841aa1c740bbaa03868dc0a403c62099a99a42 with the 72 sibling PRs of this batch: typecheck:core, file-size, complexity, cognitive-complexity, changelog-integrity green; changed-area node:test 831/831 (0 failing) and vitest 480/482 — the two reds are autoCombo/provider-family-combos.test.ts timing out at 20s, which reproduces on the PURE release tip under the full vitest suite (and is already tracked by the Release-Green issue #13866), so it is inherited, not this batch's. Merged --admin per merge-gates §3/§4/§7.