fix(e2e-ci): serve the production build in CI instead of next dev

The first full e2e run on the parallel pipeline hit the 30-min timeout:
dev-mode cold webpack compiles per route, a __webpack_modules__ TypeError,
and dev-server memory-threshold restarts. CI now builds (NEXT_DIST_DIR
matches the webServer env) and playwright webServer runs `pnpm start`;
local spec development keeps `pnpm dev`.

.github/workflows/ci-pr.yml

@@ -261,6 +261,14 @@ jobs:
         if: steps.scope.outputs.mode != 'none'
         run: pnpm db:seed:minimal
 
+      # CI e2e runs against the production build (see playwright.config.ts
+      # webServer.command); the dist dir must match the webServer env.
+      - name: Build for e2e
+        if: steps.scope.outputs.mode != 'none'
+        run: pnpm build
+        env:
+          NEXT_DIST_DIR: .next/e2e
+
       - name: E2E tests (full suite)
         if: steps.scope.outputs.mode == 'full'
         run: pnpm test:e2e

.github/workflows/ci.yml

@@ -132,5 +132,10 @@ jobs:
       - run: pnpm migrate
       - run: pnpm db:seed:minimal
 
+      - name: Build for e2e
+        run: pnpm build
+        env:
+          NEXT_DIST_DIR: .next/e2e
+
       - name: E2E tests (full suite)
         run: pnpm test:e2e

playwright.config.ts

@@ -90,7 +90,14 @@ export default defineConfig({
     },
   ],
   webServer: {
-    command: 'pnpm dev',
+    /*
+     * CI serves the PRODUCTION build (`pnpm start` after a `pnpm build` step in
+     * the e2e job): dev-mode cold webpack compiles per route plus dev-server
+     * memory restarts made the full suite blow the 30-min job timeout on CI
+     * (measured 2026-07-30 — first full e2e run on the parallel pipeline).
+     * Locally `pnpm dev` stays: no build wait and hot reload while writing specs.
+     */
+    command: process.env.CI ? 'pnpm start' : 'pnpm dev',
     /*
      * Never reuse a dev server a developer may already have running against the
      * production database — always start a fresh one bound to the test DB.