Alarm pipelines fail (problem querying data?)

[adambrown1]

After some constellation of stopping / restarting (but nothing crazy) alarm pipelines sometimes fail with endless error messages like this:

Pipeline alarm_P_IV_01 node source threw <class 'requests.exceptions.ConnectionError'>:
HTTPConnectionPool(host='192.168.131.2', port=8096):
Max retries exceeded with url: /query?q=SELECT+last%28value%29+FROM+pressure+WHERE+sensor%3D%27P_IV_01%27%3B&db=slowdata&org=pancake
(Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7fc0444be070>:
Failed to establish a new connection: [Errno 24] Too many open files'))

[adambrown1]

OK, this is actually related to #196 if a pipeline starts another one too many times. The alarms were concurrently running hundreds if not thousands of times.

However, why is there no alarm issued?