fix(ci): isolate e2e orchestration environment

apps/ui/scripts/playwright-stack.sh

@@ -17,6 +17,30 @@ POSTGRES_HOST="${CRANK_E2E_POSTGRES_HOST:-127.0.0.1}"
 USE_EXTERNAL_POSTGRES="${CRANK_E2E_USE_EXTERNAL_POSTGRES:-0}"
 ADMIN_EMAIL="${CRANK_E2E_ADMIN_EMAIL:-owner@crank.local}"
 ADMIN_PASSWORD="${CRANK_E2E_ADMIN_PASSWORD:-change-me-admin-password}"
+STARTUP_TIMEOUT_SECONDS="${CRANK_E2E_STARTUP_TIMEOUT_SECONDS:-120}"
+
+if [[ ! "$STARTUP_TIMEOUT_SECONDS" =~ ^[1-9][0-9]*$ ]]; then
+  echo "CRANK_E2E_STARTUP_TIMEOUT_SECONDS must be a positive integer" >&2
+  exit 1
+fi
+
+# These variables configure only this orchestration script. The application
+# rejects unknown variables in the owned CRANK_* namespace, so never leak the
+# E2E control plane into production process configuration.
+unset \
+  CRANK_E2E_ADMIN_EMAIL \
+  CRANK_E2E_ADMIN_PASSWORD \
+  CRANK_E2E_ADMIN_PORT \
+  CRANK_E2E_MCP_PORT \
+  CRANK_E2E_POSTGRES_DB \
+  CRANK_E2E_POSTGRES_HOST \
+  CRANK_E2E_POSTGRES_PASSWORD \
+  CRANK_E2E_POSTGRES_PORT \
+  CRANK_E2E_POSTGRES_USER \
+  CRANK_E2E_STARTUP_TIMEOUT_SECONDS \
+  CRANK_E2E_STREAM_FIXTURE_PORT \
+  CRANK_E2E_UI_PORT \
+  CRANK_E2E_USE_EXTERNAL_POSTGRES
 
 if [[ -f "$HOME/.cargo/env" ]]; then
   . "$HOME/.cargo/env"
@@ -32,6 +56,7 @@ if [[ -z "$PYTHON_BIN" ]]; then
 fi
 
 mkdir -p "$LOG_DIR"
+find "$LOG_DIR" -maxdepth 1 -type f -name '*.log' -delete
 
 (
   cd "$ROOT_DIR/apps/ui"
@@ -77,10 +102,23 @@ trap 'exit 130' INT TERM
 
 cleanup
 
+show_startup_log() {
+  local log_file="$1"
+  if [[ -f "$log_file" ]]; then
+    echo "---- $log_file ----" >&2
+    tail -n 120 "$log_file" >&2
+  fi
+}
+
 wait_for_port() {
-  local host="$1"
+  local label="$1"
-  local port="$2"
+  local host="$2"
-  until "$PYTHON_BIN" - "$host" "$port" <<'PY'
+  local port="$3"
+  local deadline=$((SECONDS + STARTUP_TIMEOUT_SECONDS))
+
+  echo "Waiting for $label on $host:$port"
+  while true; do
+    if "$PYTHON_BIN" - "$host" "$port" <<'PY'
 import socket, sys
 sock = socket.socket()
 sock.settimeout(0.5)
@@ -91,13 +129,75 @@ except OSError:
 finally:
     sock.close()
 PY
-  do
+    then
+      return 0
+    fi
+    if (( SECONDS >= deadline )); then
+      echo "Timed out waiting for $label on $host:$port after ${STARTUP_TIMEOUT_SECONDS}s" >&2
+      return 1
+    fi
+    sleep 1
+  done
+}
+
+wait_for_http() {
+  local label="$1"
+  local url="$2"
+  local pid_file="$3"
+  local log_file="$4"
+  local deadline=$((SECONDS + STARTUP_TIMEOUT_SECONDS))
+
+  echo "Waiting for $label at $url"
+  while true; do
+    if "$PYTHON_BIN" - "$url" <<'PY'
+import sys, urllib.request
+try:
+    with urllib.request.urlopen(sys.argv[1], timeout=0.5) as response:
+        sys.exit(0 if 200 <= response.status < 400 else 1)
+except Exception:
+    sys.exit(1)
+PY
+    then
+      return 0
+    fi
+
+    local pid=""
+    if [[ -f "$pid_file" ]]; then
+      pid="$(cat "$pid_file")"
+    fi
+    if [[ -z "$pid" ]] || ! kill -0 "$pid" >/dev/null 2>&1; then
+      echo "$label exited before becoming ready at $url" >&2
+      show_startup_log "$log_file"
+      return 1
+    fi
+    if (( SECONDS >= deadline )); then
+      echo "Timed out waiting for $label at $url after ${STARTUP_TIMEOUT_SECONDS}s" >&2
+      show_startup_log "$log_file"
+      return 1
+    fi
+    sleep 1
+  done
+}
+
+wait_for_postgres_container() {
+  local deadline=$((SECONDS + STARTUP_TIMEOUT_SECONDS))
+  echo "Waiting for PostgreSQL container $POSTGRES_CONTAINER"
+  while ! docker exec "$POSTGRES_CONTAINER" pg_isready -U "$POSTGRES_USER" -d "$POSTGRES_DB" >/dev/null 2>&1; do
+    if ! docker inspect "$POSTGRES_CONTAINER" >/dev/null 2>&1; then
+      echo "PostgreSQL container $POSTGRES_CONTAINER exited before becoming ready" >&2
+      return 1
+    fi
+    if (( SECONDS >= deadline )); then
+      echo "Timed out waiting for PostgreSQL container after ${STARTUP_TIMEOUT_SECONDS}s" >&2
+      docker logs "$POSTGRES_CONTAINER" >&2 || true
+      return 1
+    fi
     sleep 1
   done
 }
 
 if [[ "$USE_EXTERNAL_POSTGRES" = "1" ]]; then
-  wait_for_port "$POSTGRES_HOST" "$POSTGRES_PORT"
+  wait_for_port "PostgreSQL" "$POSTGRES_HOST" "$POSTGRES_PORT" || exit 1
 else
   docker run -d --rm \
     --name "$POSTGRES_CONTAINER" \
@@ -107,9 +207,7 @@ else
     -p "$POSTGRES_PORT:5432" \
     postgres:16-alpine >/dev/null
 
-  until docker exec "$POSTGRES_CONTAINER" pg_isready -U "$POSTGRES_USER" -d "$POSTGRES_DB" >/dev/null 2>&1; do
+  wait_for_postgres_container || exit 1
-    sleep 1
-  done
 fi
 
 export POSTGRES_HOST
@@ -141,40 +239,58 @@ chmod 700 "$CRANK_STORAGE_ROOT"
 ) &
 echo $! > "$TMP_DIR/http-fixture.pid"
 
-until curl -fsS "http://127.0.0.1:$STREAM_FIXTURE_PORT/health" >/dev/null 2>&1; do
+wait_for_http \
-  sleep 1
+  "Playwright HTTP fixture" \
-done
+  "http://127.0.0.1:$STREAM_FIXTURE_PORT/health" \
+  "$TMP_DIR/http-fixture.pid" \
+  "$LOG_DIR/http-fixture.log" || exit 1
 
-(
+if ! (
   cd "$ROOT_DIR"
   cargo run -p admin-api --bin crank-migrate -- apply >"$LOG_DIR/migrate.log" 2>&1
-)
+); then
+  echo "Failed to apply E2E database migrations" >&2
+  show_startup_log "$LOG_DIR/migrate.log"
+  exit 1
+fi
 
-BOOTSTRAP_JSON="$(
+if ! (
   cd "$ROOT_DIR"
   cargo run -p admin-api --bin crank-migrate -- admin-auth bootstrap-create \
     --email "$ADMIN_EMAIL" \
     --display-name "Crank E2E" \
     >"$LOG_DIR/bootstrap-create.log" 2>&1
-  tail -n 1 "$LOG_DIR/bootstrap-create.log"
+); then
-)"
+  echo "Failed to create the E2E bootstrap contract" >&2
-BOOTSTRAP_TOKEN="$("$PYTHON_BIN" - <<'PY' "$BOOTSTRAP_JSON"
+  show_startup_log "$LOG_DIR/bootstrap-create.log"
+  exit 1
+fi
+BOOTSTRAP_JSON="$(tail -n 1 "$LOG_DIR/bootstrap-create.log")"
+if ! BOOTSTRAP_TOKEN="$("$PYTHON_BIN" - <<'PY' "$BOOTSTRAP_JSON"
 import json, sys
 print(json.loads(sys.argv[1])["bootstrap_token"])
 PY
-)"
+)"; then
+  echo "Failed to parse the E2E bootstrap token" >&2
+  show_startup_log "$LOG_DIR/bootstrap-create.log"
+  exit 1
+fi
 printf '%s' "$BOOTSTRAP_TOKEN" >"$TMP_DIR/bootstrap-token.txt"
 printf '%s' "$ADMIN_PASSWORD" >"$TMP_DIR/admin-password.txt"
 printf '%s' "$CRANK_PASSWORD_PEPPER" >"$TMP_DIR/password-pepper.txt"
 
-(
+if ! (
   cd "$ROOT_DIR"
   cargo run -p admin-api --bin crank-migrate -- admin-auth bootstrap-complete \
     --token-file "$TMP_DIR/bootstrap-token.txt" \
     --password-file "$TMP_DIR/admin-password.txt" \
     --password-pepper-file "$TMP_DIR/password-pepper.txt" \
     >"$LOG_DIR/bootstrap-complete.log" 2>&1
-)
+); then
+  echo "Failed to complete the E2E admin bootstrap" >&2
+  show_startup_log "$LOG_DIR/bootstrap-complete.log"
+  exit 1
+fi
 
 (
   cd "$ROOT_DIR"
@@ -183,9 +299,11 @@ printf '%s' "$CRANK_PASSWORD_PEPPER" >"$TMP_DIR/password-pepper.txt"
 ) &
 echo $! > "$TMP_DIR/admin-api.pid"
 
-until curl -fsS "http://127.0.0.1:$ADMIN_PORT/health" >/dev/null 2>&1; do
+wait_for_http \
-  sleep 1
+  "admin-api" \
-done
+  "http://127.0.0.1:$ADMIN_PORT/health" \
+  "$TMP_DIR/admin-api.pid" \
+  "$LOG_DIR/admin-api.log" || exit 1
 
 (
   cd "$ROOT_DIR"
@@ -196,9 +314,11 @@ done
 ) &
 echo $! > "$TMP_DIR/mcp-server.pid"
 
-until curl -fsS "http://127.0.0.1:$MCP_PORT/health" >/dev/null 2>&1; do
+wait_for_http \
-  sleep 1
+  "mcp-server" \
-done
+  "http://127.0.0.1:$MCP_PORT/health" \
+  "$TMP_DIR/mcp-server.pid" \
+  "$LOG_DIR/mcp-server.log" || exit 1
 
 (
   cd "$ROOT_DIR/apps/ui"
@@ -206,9 +326,11 @@ done
 ) &
 echo $! > "$TMP_DIR/ui-server.pid"
 
-until curl -fsS "http://127.0.0.1:$UI_PORT/login" >/dev/null 2>&1; do
+wait_for_http \
-  sleep 1
+  "Playwright UI server" \
-done
+  "http://127.0.0.1:$UI_PORT/login" \
+  "$TMP_DIR/ui-server.pid" \
+  "$LOG_DIR/ui-server.log" || exit 1
 
 echo "Crank UI e2e stack is ready on http://127.0.0.1:$UI_PORT"