fix(healthcheck): Let Docker enforce HEALTHCHECK_TIMEOUT in check_http.py

healthcheck/check_http.py

@@ -17,28 +17,20 @@
 import urllib.error
 import urllib.request
 
-TIMEOUT = 2  # seconds
-
 
 def main(url: str) -> int:
     try:
         # Ignore HTTP(S)_PROXY, which Docker may inject into containers: the
         # target is always the container itself.
         opener = urllib.request.build_opener(urllib.request.ProxyHandler({}))
-        body = opener.open(url, timeout=TIMEOUT).read().decode()
+        body = opener.open(url).read().decode()
     except urllib.error.HTTPError as exc:
         print(f"HTTP {exc.code} from {url}", file=sys.stderr)
         return 1
     except urllib.error.URLError as exc:
         # urlopen() wraps connection-phase failures (refused, DNS, etc.) here.
         print(f"{url} unreachable: {exc.reason}", file=sys.stderr)
         return 1
-    except TimeoutError:
-        # A timeout firing during .read() (after urlopen returns) bubbles up
-        # as a bare TimeoutError from the underlying socket — not wrapped in
-        # URLError. Catch it explicitly so the message stays one-line.
-        print(f"timed out reading {url} after {TIMEOUT}s", file=sys.stderr)
-        return 1
     except OSError as exc:
         # ConnectionResetError, etc. — anything else from the socket layer.
         print(f"error against {url}: {exc}", file=sys.stderr)