Rate limit public registration writes

src/user_engine/web.py

@@ -17,6 +17,9 @@ import hmac
 import json
 import re
 import secrets
+from collections import deque
+from threading import Lock
+from time import monotonic
 from typing import Any, Callable, Iterable, Mapping
 from urllib.parse import parse_qs, urlencode, urlsplit
 
@@ -75,6 +78,8 @@ class PortalApplication:
         registration_tenants: tuple[str, ...] = (),
         registration_oidc_issuer: str = "",
         registration_password_setup_origins: tuple[str, ...] = (),
+        registration_rate_limit: int = 10,
+        registration_rate_window_seconds: int = 60,
     ) -> None:
         if len(trusted_proxy_secret) < 24:
             raise ValueError("trusted proxy secret must contain at least 24 characters")
@@ -93,6 +98,12 @@ class PortalApplication:
         self.registration_password_setup_origins = frozenset(
             origin.rstrip("/") for origin in registration_password_setup_origins
         )
+        if registration_rate_limit < 1 or registration_rate_window_seconds < 1:
+            raise ValueError("registration rate limit and window must be positive")
+        self.registration_rate_limit = registration_rate_limit
+        self.registration_rate_window_seconds = registration_rate_window_seconds
+        self._registration_attempts: dict[str, deque[float]] = {}
+        self._registration_attempts_lock = Lock()
 
     def __call__(self, environ: Mapping[str, Any], start_response: StartResponse) -> Iterable[bytes]:
         correlation_id = environ.get("HTTP_X_REQUEST_ID") or f"corr_{secrets.token_hex(12)}"
@@ -120,6 +131,16 @@ class PortalApplication:
     def _dispatch(self, environ: Mapping[str, Any], start_response: StartResponse, correlation_id: str) -> Iterable[bytes]:
         method = str(environ.get("REQUEST_METHOD", "GET")).upper()
         path = str(environ.get("PATH_INFO", "/")).rstrip("/") or "/"
+        if method == "POST" and path in {
+            "/register", "/registration/verify", "/registration/resume",
+            "/api/v1/public/registrations",
+            "/api/v1/public/registrations/verify",
+            "/api/v1/public/registrations/resume",
+        } and not self._accept_registration_attempt(environ):
+            return self._error(
+                start_response, "429 Too Many Requests", "rate_limited",
+                "Too many registration attempts. Try again later.", correlation_id,
+            )
         if path == "/healthz":
             return self._json(start_response, "200 OK", _jsonable(self.service.health()), correlation_id)
         if path == "/readyz":
@@ -1168,6 +1189,32 @@ class PortalApplication:
         if not supplied or not expected or not secrets.compare_digest(supplied, expected):
             raise AuthorizationDenied("invalid registration CSRF token")
 
+    def _accept_registration_attempt(self, environ: Mapping[str, Any]) -> bool:
+        """Apply a bounded per-source sliding-window limit to public writes.
+
+        Only ``REMOTE_ADDR`` is trusted. Forwarded headers are deliberately
+        ignored because the ingress must normalize the peer address before the
+        request reaches this process.
+        """
+        source = str(environ.get("REMOTE_ADDR") or "unknown")[:128]
+        now = monotonic()
+        cutoff = now - self.registration_rate_window_seconds
+        with self._registration_attempts_lock:
+            attempts = self._registration_attempts.setdefault(source, deque())
+            while attempts and attempts[0] <= cutoff:
+                attempts.popleft()
+            if len(attempts) >= self.registration_rate_limit:
+                return False
+            attempts.append(now)
+            if len(self._registration_attempts) > 4096:
+                for key in tuple(self._registration_attempts):
+                    values = self._registration_attempts[key]
+                    while values and values[0] <= cutoff:
+                        values.popleft()
+                    if not values:
+                        del self._registration_attempts[key]
+            return True
+
     @staticmethod
     def _page(environ: Mapping[str, Any]) -> tuple[int, int]:
         query = parse_qs(str(environ.get("QUERY_STRING", "")))