Capture warnings and report to sentry (#5697)

By default, warnings are simply printed to stderr. This makes them
easy to miss in the log. Capture warnings and user Python logger to log
them with warning level.

Also, if the message is an instance of Exception (which it typically
is), report the warning to Sentry. This is e.g. useful for asyncio
RuntimeWarning warnings "coroutine was never awaited".

supervisor/bootstrap.py

@@ -4,8 +4,10 @@
 import logging
 import os
 import signal
+import warnings
 
 from colorlog import ColoredFormatter
+from sentry_sdk import capture_exception
 
 from .addons.manager import AddonManager
 from .api import RestAPI
@@ -222,6 +224,13 @@ def initialize_system(coresys: CoreSys) -> None:
         config.path_addon_configs.mkdir()
 
 
+def warning_handler(message, category, filename, lineno, file=None, line=None):
+    """Warning handler which logs warnings using the logging module."""
+    _LOGGER.warning("%s:%s: %s: %s", filename, lineno, category.__name__, message)
+    if isinstance(message, Exception):
+        capture_exception(message)
+
+
 def initialize_logging() -> None:
     """Initialize the logging."""
     logging.basicConfig(level=logging.INFO)
@@ -248,6 +257,7 @@ def initialize_logging() -> None:
             },
         )
     )
+    warnings.showwarning = warning_handler
 
 
 def check_environment() -> None: