fix(e2e): run Playwright single-worker to stop OOM page-crash in verify-stack (refs #115)

verify-stack intermittently failed on the e2e with "Page crashed" mid-action:
Playwright defaulted to 2 workers, so two full chromium browsers ran alongside the
whole compose stack on the 8 GB runner and the renderer was OOM-killed. Pin
workers: 1 (only two long-running happy-path specs) and add --disable-dev-shm-usage,
removing the memory contention rather than masking it with retries (CLAUDE.md §15).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

BACKLOG.md

@@ -239,6 +239,12 @@ Split (issue #11 closed) into two independently-demoable slices per §13 — the
 
 **Outcome:** Boundary timer on beoordeling user task — 14 days. On timeout, reassigns to a teamlead role.
 
+### S-26 · Self-service — resume an existing registration after refresh — #111
+
+**Outcome:** a signed-in zorgprofessional who reloads the self-service portal (or returns later) gets back to their in-flight registration and its actions (Documenten aanleveren, Trek aanvraag in), instead of a blank submit form with the reference lost. Today all post-submit state lives in in-memory signals, the reference is not in the URL, and there is no self-service read endpoint — so a reload strands the registration. Adds an owner-scoped (DigiD bsn) `GET /self-service/registrations` on the BFF/domain and a load-on-init/route restore in the portal.
+
+**Acceptance:** BDD — resume after refresh shows the existing registration; lookup is owner-scoped (never another citizen's); a user with no in-flight registration still sees the submit form. Playwright e2e reloads mid-flow and asserts the actions remain reachable.
+
 ---
 
 ## Iteration 3 — Maintenance portal and observability *(milestone: `Iteration 3 — Beheer & Observability`)*

tests/e2e/playwright.config.ts

@@ -12,6 +12,12 @@ export default defineConfig({
   timeout: 90_000,
   expect: { timeout: 15_000 },
   retries: 1,
+  // Run the specs serially. Each spec drives a full `channel: 'chromium'` browser, and the e2e
+  // shares an 8 GB runner with the entire compose stack (OpenZaak, NRC, Keycloak, Flowable, 4×
+  // Postgres, every service + 3 portals). Two parallel browsers exhaust memory and the renderer is
+  // OOM-killed mid-action ("Page crashed") — fixing the flakiness at its source rather than leaning
+  // on `retries` (CLAUDE.md §15). Only two long-running happy-path specs, so serial costs little.
+  workers: 1,
   reporter: [['list']],
   use: {
     baseURL,
@@ -26,7 +32,13 @@ export default defineConfig({
     // headless), not Playwright's default headless-shell, so pin `channel: 'chromium'`.
     channel: 'chromium',
     launchOptions: {
-      args: [`--unsafely-treat-insecure-origin-as-secure=${baseURL},${behandelURL}`],
+      args: [
+        `--unsafely-treat-insecure-origin-as-secure=${baseURL},${behandelURL}`,
+        // Write Chromium's shared memory to /tmp instead of the container's small /dev/shm, so a
+        // large DOM/heap can't crash the renderer on the memory-constrained runner (belt-and-braces
+        // alongside the single worker above).
+        '--disable-dev-shm-usage',
+      ],
     },
   },
   projects: [{ name: 'chromium', use: { ...devices['Desktop Chrome'] } }],