Act on the wj architectural review: races, limits, pool and contract gaps

Findings F2, F4, F5, F6, F7, F8, F9, F10, F11 and F13 from
wj-architectural-review.md. F1/F1a (the non-sargable LOWER pattern) and F3
(the warranty sync holding a pooled connection) are being handled separately
and are untouched here.

- A collector check-in that raced another for the same PC lost everything:
  the unique index on the application fingerprint raised, nothing caught it,
  and the whole check-in rolled back, discarding state, events and hardware,
  answering the agent with form copy meant for a person. The insert now runs
  in a savepoint, reconciles what the other transaction already wrote, one row
  per savepoint so a second collision skips a row rather than the batch, and
  the check-in commits with a warning. This is the shape the delete side of
  the same function already had.
- Enrollment is deliberately unauthenticated, so it now carries a per-address
  rate limit and a cap on pending requests, and the cap is a setting an admin
  can see and change. A flood can no longer bury the approval queue or grow
  the table without bound.
- The fleet alert poll no longer formats debug strings on every request.
- BusinessUnit is on the contract surface (0.34.0), so the endpoints plugin
  stops reaching into a core table with raw SQL.
- A test walks every asset type descriptor and proves the endpoint it names
  really refuses a caller without the declared permission.
- The plugin settings routes now need a permission; they were telling anyone
  which modules a site runs and where its components live.
- The connection pool is sized explicitly with a short timeout, and the
  arithmetic against the server's connection ceiling is written next to the
  worker count instead of being left to defaults nobody chose.
- Fleet alerts carry both the correct lowercase keys and the old camelCase
  pair. Trays older than 1.7.0 match keys case-sensitively and would have gone
  quiet on every PC not yet upgraded; the old pair goes when they are.
- The dev server proxy points at localhost again, with environment variables
  for the remote-box workflow, and one developer's address is out of the
  server certificate page.

plugins/endpoints/api/ratelimit.py

@@ -0,0 +1,52 @@
+"""Fixed-window per-IP rate limiting for the agent-facing routes.
+
+One implementation, several budgets. The bootstrap installer routes and the
+enrollment route are both unauthenticated by design - a PC with no agent has
+neither a credential nor a user session - so the only thing standing between
+them and a loop is a limiter. Each caller gets its own budget name so the two
+surfaces cannot spend each other's allowance.
+
+Same shape and same cache extension as the login limiter in
+shopdb.core.api.auth. Under the default SimpleCache the counter is per-process,
+so with N gunicorn workers the effective budget is N times the configured
+maximum; that is defence in depth rather than a hard quota, and a shared cache
+backend (Redis/memcached) tightens it to a true global budget.
+"""
+
+import time
+
+from flask import current_app, request
+
+from shopdb.api import cache, client_ip, error_response
+
+
+def caller_ip():
+    return client_ip(request) or 'unknown'
+
+
+def ratelimited(budget, configprefix, defaultmax, defaultwindow):
+    """True when this caller is over budget for the current window.
+
+    budget        cache key namespace, e.g. 'endpointenrollratelimit'
+    configprefix  config key stem, e.g. 'ENDPOINTS_ENROLL_RATELIMIT'; reads
+                  <prefix>_ENABLED, <prefix>_MAX and <prefix>_WINDOW_SECONDS
+    """
+    if not current_app.config.get(configprefix + '_ENABLED', True):
+        return False
+    window = current_app.config.get(configprefix + '_WINDOW_SECONDS',
+                                    defaultwindow)
+    maxhits = current_app.config.get(configprefix + '_MAX', defaultmax)
+    # Time bucket makes this a fixed window: the key rolls over at each window
+    # boundary, so a per-hit set() cannot turn it into a sliding window.
+    bucket = int(time.time() // window) if window > 0 else 0
+    key = '{0}:{1}:{2}'.format(budget, caller_ip(), bucket)
+    count = cache.get(key) or 0
+    if count >= maxhits:
+        return True
+    cache.set(key, count + 1, timeout=window)
+    return False
+
+
+def rate_limit_response(message):
+    """The standard 429 shape every limiter in the product answers with."""
+    return error_response('RATE_LIMITED', message, http_code=429)