[BUG]: CSM Observability documentation is complicated and causing confusion

[gallacher]

Bug Description

1. The second step (from link) in the Helm based install is copying the secret from the CSI namespace to the CSM namespace … it does “Note” that the namespace for the CSM must exist prior to this step being taken; my recommendation is to make sure this step is indeed the last step (visually) in this Helm deployment as putting it second (past the cert-manager install) is confusing.

2. Topology Module
   Source: https://dell.github.io/csm-docs/docs/observability/

Note: By going to that link it does point you to: “For documentation, please visit Container Storage Modules documentation.” Which as we can see there is no Topology

“Please visit the repository for more information” should state, Please visit the repository for installation instruction. Because, there is nothing that explicitly states you have to go through a deployment of the Topology Module. The only semblance of this if you catch it, the first line in the values.yaml when deploying the CSM-OBS module calls out the csm-topology:v1.0 container….

Or, we provide a precise set of deployment instructions as part of our overall documentation. You’ll notice those instructions are not present below, which is a screen shot from the following landing page for CSM-docs. Thus, it can be implied or assumed by the installing party that they will benefit from Topology as a byproduct of the CSM OBS module being installed…?

3. Installing Prometheus and Grafana using the Operator post Observability deployment

This does start to escape our realm of responsibilities as we are not the purveyors of either Prometheus or Grafana, however, what we need to do is precisely state how to deploy the scrape for Prometheus when deployed by an Operator in OpenShift. Or, should a customer already have Prometheus and Grafana deployed for other needs in their environment, we must be able to provide details on injecting that scrape. Where, our documentation provides the recommendation of deploying Prometheus and Grafana after CSM Observability is installed.

Once Prometheus was deployed, there was plenty of fumbling to introduce the scrape. Admittedly, this is not something I’d done before but I’ve been over and over the documentation, other notes, etc. But the fact of the matter is this, when things are deployed via an Operator, there is (are) very specific methods of altering the state of that plugin. Whereas, the manual method (which we ended up using) in deploying Prometheus is much easier to navigate.

4. Simplified steps for CSM Observability deployment

Here are some more simplified steps for CSM Observability deployment that would be helpful to reflect in the documentation.

Simple Steps.txt

Logs

N/A

Screenshots

N/A

Additional Environment Information

N/A

Steps to Reproduce

No steps to reproduce. Documentation related

Expected Behavior

Documentation should be updated to align with changes outlined in the description.

CSM Driver(s)

CSI Driver for PowerFlex v2.1.0

Installation Type

Helm

Container Storage Modules Enabled

Observability v1.1.0

Container Orchestrator

OpenShift

Operating System

N/A

[shaynafinocchiaro]

Hello @OA72280, I've reviewed your suggestions for the CSM Observability documentation. Regarding the Helm deployment steps, a successful installation of the helm chart is dependent on the CSI driver config Secrets being in the Observability namespace. So, that step has to remain a step before installing the chart. We can however reformat the Helm deployment page and make it easier to follow.