Running installed nrpe check fails

[Michael Nelson]

During nrpe-external-master-relation-changed a nagios check is configured and installed, but running that nagios check fails with:

ERROR: Could not connect to RabbitMQ server localhost:5672
Exception socket.error: (107, 'Transport endpoint is not connected') in <bound method TCPTransport.__del__ of <amqplib.client_0_8.transport.TCPTransport object at 0x7f33bd9685d0>> ignored

Steps to repro:
 * deploy rabbitmq charm (on trusty?)
 * add nrpe-external-master subordinate and relate
 * find the nagios check: `cat /etc/nagios/nrpe.d/check_rabbitmq.cfg`, and run it manually.

Result: above error
Expected result: same

Notes:
 * Using the guest vhost (which doesn't require user/pass) works:

$ /usr/local/lib/nagios/plugins/check_rabbitmq.py --vhost /
Ok: sent and received 10 test messages

[David Ames (thedac)]

A nagios user/pass is used in the config now. Confirmed this works likely since 15.04

[Paul Gear (paulgear)]

We experienced these symptoms after an upgrade from 15.04 charms to 17.11 charms. This caused our previously-working rabbitmq nagios check to stop working. Log extract: https://pastebin.canonical.com/p/Mp7HHBGR9K/

Resetting the rabbitmq password using "rabbitmqctl change_password nagios-rabbitmq-server-1 $(cat /var/lib/charm/rabbitmq-server/nagios-rabbitmq-server-1.local_passwd)" fixed the problem.

@thedac: Should we be creating a new bug for this?

[David Ames (thedac)]

I believe it is safe to age this bug out. The original bug was fixed per comment #1.

Work has been done on the rabbitmq charm in regards to NRPE monitoring in the intervening time frame. If Paul believes otherwise, we should probably open a new bug.