fix: remove failing key check from sentry logging

Twisteds new logging does not use the 'failure' key at all, the
prior assumptions it would exist caused key exceptions instead
of reporting errors. This fix changes to only the new style of
failure checking.

Issue #493

autopush/logging.py

@@ -78,6 +78,7 @@ def __init__(self, logger_name, log_level="debug", log_format="json",
             self.raven_client = raven.Client(
                 release=raven.fetch_package_version("autopush"),
                 transport=TwistedHTTPTransport,
+                enable_breadcrumbs=False,
             )
         else:
             self.raven_client = None
@@ -88,9 +89,8 @@ def __init__(self, logger_name, log_level="debug", log_format="json",
             self.firehose = None
 
     def __call__(self, event):
-        if self.raven_client and ('failure' in event or
-                                  'log_failure' in event):
-            f = event.get("log_failure", event["failure"])
+        if self.raven_client and 'log_failure' in event:
+            f = event["log_failure"]
             reactor.callFromThread(
                 self.raven_client.captureException,
                 exc_info=(f.type, f.value, f.getTracebackObject())
@@ -109,7 +109,7 @@ def __call__(self, event):
             self._output.flush()
 
     def json_format(self, event):
-        error = bool(event.get("isError")) or "failure" in event
+        error = bool(event.get("isError")) or "log_failure" in event
         ts = event["log_time"]
 
         if error: