Used time.monotonic() instead of time.time() where applicable.

time.monotonic() available from Python 3.3:

- Nicely communicates a narrow intent of "get a local system monotonic
  clock time" instead of possible "get a not necessarily accurate Unix
  time stamp because it needs to be communicated to outside of this
  process/machine" when time.time() is used.
  
- Its result isn't affected by the system clock updates.

There are two classes of time.time() uses changed to time.monotonic()
by this change:

- measuring time taken to run some code.

- setting and checking a "close_at" threshold for for persistent db
  connections (django/db/backends/base/base.py).

1 files changed, 2 insertions, 2 deletions

diff --git a/docs/topics/db/instrumentation.txt b/docs/topics/db/instrumentation.txt
index 529347094f..933c67a96d 100644
--- a/docs/topics/db/instrumentation.txt
+++ b/docs/topics/db/instrumentation.txt
@@ -69,7 +69,7 @@ For a more complete example, a query logger could look like this::
 
         def __call__(self, execute, sql, params, many, context):
             current_query = {'sql': sql, 'params': params, 'many': many}
-            start = time.time()
+            start = time.monotonic()
             try:
                 result = execute(sql, params, many, context)
             except Exception as e:
@@ -80,7 +80,7 @@ For a more complete example, a query logger could look like this::
                 current_query['status'] = 'ok'
                 return result
             finally:
-                duration = time.time() - start
+                duration = time.monotonic() - start
                 current_query['duration'] = duration
                 self.queries.append(current_query)