get_running_builds very slow to return results

[Bao Haojun]

In our site, I have found get_running_builds to be very slow.

I have found that it's because depth=2 is used when calling get_node_info, and it seems to be fixed by the following patch:

    diff --git a/jenkins/__init__.py b/jenkins/__init__.py
    index b443756..20cc17c 100755
    --- a/jenkins/__init__.py
    +++ b/jenkins/__init__.py
    @@ -1411,7 +1411,7 @@ class Jenkins(object):
                 else:
                     node_name = node['name']
                 try:
    - info = self.get_node_info(node_name, depth=2)
    + info = self.get_node_info(node_name, depth=1)
                 except JenkinsException as e:
                     # Jenkins may 500 on depth >0. If the node info comes back
                     # at depth 0 treat it as a node not running any jobs.

After we change the depth to 1, I can still get the list of running builds, so depth=2 seems to be a waste. But I may be missing something here, is there a reason why we must use depth=2 instead of depth=1?

[Bao Haojun (baohaojun)]

When using depth=2, it takes 6 minutes to return, after changing it to depth=1, it takes 2 seconds only.