Fix inventory log timestamp, organize logging code

The fixes and issue where the timestaps in the stdout for
inventory updates gave the time since the start of the dispatcher
instead of the time since the start of the update.

This commit also moves the handler into the utils module where
other custom AWX handlers live, instead of tasks.py
this is to keep tasks.py relatively clean, as best as possible

awx/main/utils/formatters.py

@@ -9,6 +9,7 @@ import socket
 from datetime import datetime
 
 from dateutil.tz import tzutc
+from django.utils.timezone import now
 from django.core.serializers.json import DjangoJSONEncoder
 from django.conf import settings
 
@@ -17,8 +18,15 @@ class TimeFormatter(logging.Formatter):
     '''
     Custom log formatter used for inventory imports
     '''
+    def __init__(self, start_time=None, **kwargs):
+        if start_time is None:
+            self.job_start = now()
+        else:
+            self.job_start = start_time
+        super(TimeFormatter, self).__init__(**kwargs)
+
     def format(self, record):
-        record.relativeSeconds = record.relativeCreated / 1000.0
+        record.relativeSeconds = (now() - self.job_start).total_seconds()
         return logging.Formatter.format(self, record)
 
 

awx/main/utils/handlers.py

@@ -7,6 +7,10 @@ import os.path
 
 # Django
 from django.conf import settings
+from django.utils.timezone import now
+
+# AWX
+from awx.main.exceptions import PostRunError
 
 
 class RSysLogHandler(logging.handlers.SysLogHandler):
@@ -40,6 +44,58 @@ class RSysLogHandler(logging.handlers.SysLogHandler):
             pass
 
 
+class SpecialInventoryHandler(logging.Handler):
+    """Logging handler used for the saving-to-database part of inventory updates
+    ran by the task system
+    this dispatches events directly to be processed by the callback receiver,
+    as opposed to ansible-runner
+    """
+
+    def __init__(self, event_handler, cancel_callback, job_timeout, verbosity,
+                 start_time=None, counter=0, initial_line=0, **kwargs):
+        self.event_handler = event_handler
+        self.cancel_callback = cancel_callback
+        self.job_timeout = job_timeout
+        if start_time is None:
+            self.job_start = now()
+        else:
+            self.job_start = start_time
+        self.last_check = self.job_start
+        self.counter = counter
+        self.skip_level = [logging.WARNING, logging.INFO, logging.DEBUG, 0][verbosity]
+        self._current_line = initial_line
+        super(SpecialInventoryHandler, self).__init__(**kwargs)
+
+    def emit(self, record):
+        # check cancel and timeout status regardless of log level
+        this_time = now()
+        if (this_time - self.last_check).total_seconds() > 0.5:  # cancel callback is expensive
+            self.last_check = this_time
+            if self.cancel_callback():
+                raise PostRunError('Inventory update has been canceled', status='canceled')
+        if self.job_timeout and ((this_time - self.job_start).total_seconds() > self.job_timeout):
+            raise PostRunError('Inventory update has timed out', status='canceled')
+
+        # skip logging for low severity logs
+        if record.levelno < self.skip_level:
+            return
+
+        self.counter += 1
+        msg = self.format(record)
+        n_lines = len(msg.strip().split('\n'))  # don't count line breaks at boundry of text
+        dispatch_data = dict(
+            created=now().isoformat(),
+            event='verbose',
+            counter=self.counter,
+            stdout=msg,
+            start_line=self._current_line,
+            end_line=self._current_line + n_lines
+        )
+        self._current_line += n_lines
+
+        self.event_handler(dispatch_data)
+
+
 ColorHandler = logging.StreamHandler
 
 if settings.COLOR_LOGS is True: