Add dequeue method check to add_queue

[nasark]

We need to be checking which dequeue method is being used when queueing events. This is because we don't want all events to be processed by the messaging client (kafka) and only when needed i.e. event syndication.

@miq-bot assign @agrare
@miq-bot add_label bug, quinteros/yes?

[agrare]

Since we care about the event handler specifically

Suggested change

	if MiqQueue.messaging_client('event_handler').present? && MiqQueueWorkerBase.worker_settings[:dequeue_method] != "drb"
	if MiqQueue.messaging_client('event_handler').present? && MiqEventHandler.worker_settings[:dequeue_method] != "drb"

[agrare]

This looks good, I wonder if (follow-up) we should expose this via a class method on the MiqQueueWorkerBase class since the Runner dequeue_method_via_miq_messaging? method is basically

self.class.worker_settings[:dequeue_method] == :miq_messaging && MiqQueue.messaging_type != "miq_queue" && MiqQueue.messaging_client(self.class.name).present?`

which is essentially the same logic.

[nasark]

This looks good, I wonder if (follow-up) we should expose this via a class method on the MiqQueueWorkerBase class since the Runner dequeue_method_via_miq_messaging? method is basically

self.class.worker_settings[:dequeue_method] == :miq_messaging && MiqQueue.messaging_type != "miq_queue" && MiqQueue.messaging_client(self.class.name).present?`
which is essentially the same logic.

Ah yeah that was my initial approach but I wasn't sure if exposing the runner methods was a direction we wanted to go, I'll take note for a follow-up

[agrare]

Ah yeah that was my initial approach but I wasn't sure if exposing the runner methods was a direction we wanted to go

Yeah we won't be able to get to the Runner instance from here, but both should be able to see the worker class

[nasark]

To pass specs, added dequeue method settings stub

[agrare]

Actually sorry I missed this wasn't a - on the first line, better to pass both keys to the stub_settings_merge

Suggested change

	before { stub_settings_merge(:messaging => {:type => 'artemis'}) }
	before { stub_settings_merge(:workers => {:worker_base => {:queue_worker_base => {:defaults => {:dequeue_method => "miq_messaging"}}}}) }
	before do
	stub_settings_merge(
	:messaging => {:type => 'artemis'}),
	:workers => {:worker_base => {:queue_worker_base => {:defaults => {:dequeue_method => "miq_messaging"}}}}
	)
	end

[agrare]

@nasark given the bug we had I think we should add a context here for messaging_type=>[artemis/kafka] but dequeue_method=>drb and add a test to confirm we use miq_queue not manageiq_messaging

[agrare]

Same here

[nasark]

Added specs and reordered conditional so we don't needlessly create a messaging client object

[miq-bot]

Checked commit nasark@da55e4a with ruby 2.6.10, rubocop 1.28.2, haml-lint 0.35.0, and yamllint
2 files checked, 0 offenses detected
Everything looks fine. ⭐

[bdunne]

Backported to quinteros in 1ee2795

commit 1ee2795c82ac4ad0ee4b5b06d48c3a74cb479927 (HEAD -> quinteros, upstream/quinteros)
Author: Adam Grare <[email protected]>
Date:   Wed Oct 4 13:58:23 2023 -0400

    Merge pull request #22728 from nasark/dequeue_messaging_check
    
    Add dequeue method check to add_queue
    
    (cherry picked from commit c7e045b629d65506e5c6e6b1d9d1bce419798e22)