Increase log level for command failures

It's hard to find the actual root cause of the container configurations
when looking in the logs because the error messages are just WARNINGs.
Let's raise the messages where the real error message is to an ERROR
level.

Change-Id: Icb7138de7f6065c758cd39cb0e433c6412442b1c
Closes-Bug: #1819931

docker/container-puppet.py

@@ -490,9 +490,9 @@ def mp_puppet_config(*args):
                 rm_container(uname)
                 break
             time.sleep(3)
-            log.warning('%s run failed after %s attempt(s): %s' % (cmd,
-                                                                   cmd_stderr,
-                                                                   count))
+            log.error('%s run failed after %s attempt(s): %s' % (cmd,
+                                                                 cmd_stderr,
+                                                                 count))
             log.warning('Retrying running container: %s' % config_volume)
         else:
             if cmd_stdout: