diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 214e293..0cf6048 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -88,10 +88,22 @@ jobs:
           supabase status -o env | grep -E '^(API_URL|PUBLISHABLE_KEY)=' > /tmp/sb.env
           . /tmp/sb.env
           printf 'VITE_SUPABASE_URL=%s\nVITE_SUPABASE_PUBLISHABLE_KEY=%s\nVITE_APP_URL=http://localhost:4173\n' "$API_URL" "$PUBLISHABLE_KEY" > .env.local
+      # The local Realtime tenant allows 100 Postgres change messages a second and drops the rest
+      # (MessagePerSecondRateLimitReached), which parallel journeys pass. Raise it, then restart
+      # Realtime so it rereads the tenant.
+      - name: Raise Realtime limits
+        run: |
+          docker exec supabase_db_potluck psql -U supabase_admin -d postgres -v ON_ERROR_STOP=1 -c             "update _realtime.tenants set max_events_per_second = 1000, max_joins_per_second = 500 where external_id = 'realtime-dev'"
+          docker restart supabase_realtime_potluck
+          for i in $(seq 1 30); do
+            [ "$(docker inspect -f '{{.State.Health.Status}}' supabase_realtime_potluck)" = healthy ] && break
+            sleep 2
+          done
+          docker inspect -f '{{.State.Health.Status}}' supabase_realtime_potluck
       - run: pnpm --filter @potluck/web exec playwright install --with-deps chromium webkit
       - run: pnpm e2e
-      # Live updates are silent when Realtime drops a change (a policy that errors or times out
-      # for the subscriber), so keep its log when a journey fails.
+      # Live updates are silent when Realtime drops a change (a rate limit, or a policy that errors
+      # or times out for the subscriber), so keep its log when a journey fails.
       - name: Realtime log
         if: failure()
         run: docker logs supabase_realtime_potluck 2>&1 | tail -n 400
diff --git a/docs/testing.md b/docs/testing.md
index 1ad7494..2a93b13 100644
--- a/docs/testing.md
+++ b/docs/testing.md
@@ -130,7 +130,7 @@ pnpm --filter @potluck/web exec playwright test e2e/expenses.spec.ts --project=d
 pnpm --filter @potluck/web exec playwright show-report               # after a failure
 ```
 
-Locally, a server already listening on :4173 is reused, even if another checkout or worktree started it, so the tests may run another branch's build. If you change app code while one is running, rebuild with `VITE_APP_URL=http://localhost:4173 pnpm --filter @potluck/web build` or stop it first. `E2E_PORT=4174` runs on another port beside it. `E2E_WORKERS` sets the local worker count (default 2).
+Locally, a server already listening on :4173 is reused, even if another checkout or worktree started it, so the tests may run another branch's build. If you change app code while one is running, rebuild with `VITE_APP_URL=http://localhost:4173 pnpm --filter @potluck/web build` or stop it first. `E2E_PORT=4174` runs on another port beside it. `E2E_WORKERS` sets the local worker count (default 2). The local Realtime tenant drops Postgres changes past 100 a second (`MessagePerSecondRateLimitReached` in `docker logs supabase_realtime_potluck`), which looks like a live update that never arrives. CI raises it in the e2e job; locally, run the same `update _realtime.tenants` from `ci.yml` (as `supabase_admin`; `postgres` may not write it) and restart the Realtime container if live journeys fail only under load.
 
 Projects (`apps/web/playwright.config.ts`):