ci: balance remaining server suites with measured native Runner cost

Co-Authored-By: Paperclip <noreply@paperclip.ing>

scripts/__tests__/release-verify-workflow.test.mjs

@@ -209,7 +209,7 @@ test("release verify workflow covers the same split test surface as stable PR ve
   for (const group of ["general-server-without-chat", "general-chat", "general-workspaces-a", "general-workspaces-b"]) {
     assert.match(verifyWorkflow, new RegExp(`group: ${group}`));
   }
-  for (const [group, count] of [["general-server-without-chat", 4], ["general-chat", 3]]) {
+  for (const [group, count] of [["general-server-without-chat", 5], ["general-chat", 3]]) {
     const rows = [...verifyWorkflow.matchAll(new RegExp(`group: ${group}\\n\\s+group_label: [^\\n]+\\n\\s+shard_index: (\\d+)\\n\\s+shard_count: (\\d+)`, "g"))];
     assert.deepEqual(rows.map((row) => [Number(row[1]), Number(row[2])]),
       Array.from({ length: count }, (_, index) => [index, count]));

scripts/__tests__/run-vitest-stable-shard.test.mjs

@@ -270,9 +270,9 @@ test("the real shard partition is duration-balanced", () => {
 
 test("release server shards plus the dedicated chat file cover the original server group exactly", () => {
   const full = dryRunJson(["--mode", "general", "--group", "general-server", "--shard-index", "0", "--shard-count", "1"]);
-  const shards = Array.from({ length: 4 }, (_, index) => dryRunJson([
+  const shards = Array.from({ length: 5 }, (_, index) => dryRunJson([
     "--mode", "general", "--group", "general-server-without-chat",
-    "--shard-index", String(index), "--shard-count", "4",
+    "--shard-index", String(index), "--shard-count", "5",
   ]));
   const files = shards.flatMap((shard) => shard.selectedGeneralServerSuites);
   const chat = "server/src/__tests__/chat-channels.integration.test.ts";

scripts/general-server-shard-durations.json

@@ -1,7 +1,9 @@
 {
   "$comment": "Per-suite wall-clock durations (ms) for the general-server vitest lane, used by scripts/general-server-shard.mjs to balance suites across the PR shard matrix. Sampled from a real PR run of .github/workflows/pr.yml (actions run 32708351172, 2026-08-24) by diffing consecutive per-suite completion timestamps in the 'Run grouped general test suites' logs \u2014 that captures each suite's true serial cost (import + collect + tests), not just the vitest-reported test time. Suites missing here get the median weight, so the manifest only needs occasional refreshes.",
   "$chatSample": "chat-channels.integration.test.ts: actions run 34405038082, job 102646337040, 2026-09-09. The first suite completed at 21:19:52.9026416Z after Vitest RUN at 21:09:07.5491992Z: 645354ms rounded up, including startup/import/collection; the 985 tests themselves took 629654ms. All 2972 tests in the shard passed, but the job exceeded its unchanged 20-minute bound during cleanup. Recording this missing heavy-suite weight lets the existing LPT partition reserve one of the existing five shards without changing test coverage, isolation, or deadlines.",
+  "$nativeRunnerSample": "native-codex-runner.integration.test.ts: actions run 34555686996, job 103127786254, 2026-09-11. Consecutive suite completions at 02:50:34.2830368Z and 02:55:07.9837588Z give 273701ms including import/collection (test body 270773ms). This previously unweighted suite made one four-way server shard take 14m38s; recording its cost lets the existing LPT partition balance it in both PR and release runs.",
   "durations": {
+    "server/src/services/native-runtime/native-codex-runner.integration.test.ts": 273701,
     "server/src/__tests__/access-service.test.ts": 4757,
     "server/src/__tests__/access-validators.test.ts": 645,
     "server/src/__tests__/activity-log-responsible-user.test.ts": 4407,