action check-queues does not report the size of queues

[Drew Freiberger]

When running 'juju run-action rabbitmq-server/0 check-queues' the output is a list of queues from the vhost, but this does not include the sizes of the queues as the code appears to intend at https://github.com/openstack/charm-rabbitmq-server/blob/master/actions/actions.py#L94

It would be very helpful for environments where ssh is blocked to have this additional queue depth 'size' in the returned output.

After running juju run-action rabbitmq-server/0 check-queues queue-depth=1, I get the following output:

results:
  outcome: Success
  output: '[''alarm.all.sample'']'
status: completed
timing:
  completed: 2019-07-01 19:03:52 +0000 UTC
  enqueued: 2019-07-01 19:03:41 +0000 UTC
  started: 2019-07-01 19:03:50 +0000 UTC

I believe the "list()" wrapper around the dict-comprehension may be returning only the keys of the dictionary, rather than the intended [queue, size, queue, size, queue, size...] result.

[Alvaro Uria (aluria)]

Apart from this, there should be a juju config parameter to limit the size of certain (unconsumed) queues. "notification-ttl" limits per time, but a queue could grow too much and affect rest of (consumed) queues performance.

[Robert Gildein (rgildein)]

There is another bug, because `rabbitmqctl` print table header by default now. We need to use better regex to filter the output of the `rabbitmqctl list_queues` command or add the` --no-table-headers` flag.

[Peter Sabaini (peter-sabaini)]

I think we might have a duplicate here in Bug #1838964, proposed change https://review.opendev.org/#/c/759407/

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.opendev.org/761602

[OpenStack Infra (hudson-openstack)]

Change abandoned by Robert Gildein (<email address hidden>) on branch: master
Review: https://review.opendev.org/761602
Reason: https://review.opendev.org/#/c/759407/