[KOGITO-378] Integrate Kogito Operator with Prometheus Operator

[xieshenzh]

Jira: https://issues.jboss.org/browse/KOGITO-378

Description: deploy a Service Monitor for scraping the metrics of the Kogitio app, if Prometheus Operator is installed.

Many thanks for submiting your Pull Request ❤️!

Please make sure that your PR meets the following requirements:

• You have read the contributors guide
• Pull Request title is properly formatted: [KOGITO-XYZ] Subject
• Pull Request contains link to the JIRA issue
• Pull Request contains description of the issue
• Pull Request does not include fixes for issues other than the main ticket
• Your feature/bug fix has a unit test that verifies it
• You've tested the new feature/bug fix in an actual OpenShift cluster

[ricardozanini]

@xiezhang7 thanks for your work! Just a small question before I start the actual review. Please see below. Thanks!

Also, I believe that we have some failing tests.

[sutaakar]

Briefly went through the PR, didn't find there anything suspicious.
I will make more detailed review later, possible today or tomorrow.
If I won't make it this week then don't wait with merging for me.

[ricardozanini]

@xiezhang7 is it possible to use the Discovery client to find if the Prometheus API is available in the cluster like we do today with isOpenShift method in the client package?

This way you just create the ServiceMonitor if the API is available using the CR and the standard client.

[xieshenzh]

@xiezhang7 is it possible to use the Discovery client to find if the Prometheus API is available in the cluster like we do today with isOpenShift method in the client package?

This way you just create the ServiceMonitor if the API is available using the CR and the standard client.

@ricardozanini yes, we can do that. But we will still need the prometheus client to get the list of deployed service monitors, because the standard client doesn't support that.

[ricardozanini]

@xiezhang7 is it possible to use the Discovery client to find if the Prometheus API is available in the cluster like we do today with isOpenShift method in the client package?
This way you just create the ServiceMonitor if the API is available using the CR and the standard client.

@ricardozanini yes, we can do that. But we will still need the prometheus client to get the list of deployed service monitors, because the standard client doesn't support that.

Cool! But why do we need to list the ServiceMonitors?

[xieshenzh]

Cool! But why do we need to list the ServiceMonitors?

Because operator-utils is used to maintain the resources. It lists the resources and filters them by the ownership. It doesn't make any assumption on the resource name or the amount of the resources.

[ricardozanini]

Cool! But why do we need to list the ServiceMonitors?

Because operator-utils is used to maintain the resources. It lists the resources and filters them by the ownership. It doesn't make any assumption on the resource name or the amount of the resources.

hmm I see. Well, I'm not that happy with this solution, but I see value using the operator-utils to manage our resources. It is what it is.

Could you please them rebase your commit?

[xieshenzh]

Thanks, @ricardozanini . I have rebased the code, and changed the code to use Discovery client to check the Prometheus API.

[ricardozanini]

LGTM in general, @xiezhang7. Just minor comments.

Many thanks for the hard work!

[ricardozanini]

Thanks @xiezhang7

[spolti]

I didn't find anything suspicious as well, but, as a user, it is no clear how it works, or what exactly needs to be done to enable/access the metrics, and how to take advantage of this integration, the user will need to do some manual configuration besides add the prometheus addon on their project, or will the operator take care of the prometheus operator integration and run it automatically?
Do you mind to add a few detailed information about it?

[xieshenzh]

I didn't find anything suspicious as well, but, as a user, it is no clear how it works, or what exactly needs to be done to enable/access the metrics, and how to take advantage of this integration, the user will need to do some manual configuration besides add the prometheus addon on their project, or will the operator take care of the prometheus operator integration and run it automatically?
Do you mind to add a few detailed information about it?

Thanks, @spolti . I have updated the README.

1. Accessing the metrics is configured by the prometheus operator. The integration only works when there is already a prometheus operator install in the cluster, which means the user is supposed to have some knowledge of the prometheus operator. Mainly, the user will have to create or configure a Prometheus instance for the prometheus operator. I added the information in the README .

2. Enabling/generating metrics is configured in the kogito source project. It is mentioned in the first part of the Prometheus Integration section.

3. The integration is to let the kogito operator create a Service Monitor instance of the prometheus operator automatically, when prometheus operator is installed and metrics for the kogito app is enabled. The Service Monitor instance is responsible for describing how the Prometheus instance should scrape the metrics of the Kogito App.

So the user is supposed to finish 1 and 2. The operator will do 3.