Refactor services init jobs

[imgios]

• Please check if the PR fulfills these requirements

• [x] The branch naming convention follows our guidelines

• [x] Docs have been added / updated (for bug fixes / features)

• What kind of change does this PR introduce? (Bug fix, feature, docs update, ...)

This PR brings few new things:

• Services init operations are now performed in specific job. In details, now we have one job per service, and each job runs a Python script that setup the service via REST API (as discussed in #11)
• Each job is using env variable, to make it quite flexible
• Job execution order set by using a weigth. The order is the following:
  • pre-install
  • radarr
  • sonarr
  • prowlarr
  • jellyfin
  • jellyseerr
• cURL image tag has been changed from latest to 8.7.1
• Added the ImagePullPolicy for the jobs and set it to IfNotPresent, we can think about making it a parameter from the values but we can decide later
• Added a basic request check on the response status code to let the job fail if the flow fails somewhere.

I hope I didn't forget anything.

• What is the current behavior? (You can also link to an open issue here)

We don't have any ImagePullPolicy defined and the services init are mixed in the same jobs. Moreover, almost all the images are using the latest tag instead of a given one.

• What is the new behavior (if this is a feature change)?

One job per service executed in a given order, with the ImagePullPolicy set to IfNotPresent in order to pull the image only if it's missing from the node.

• Does this PR introduce a breaking change? (What changes might users need to make in their application due to this PR?)

Not really I guess

• Other information:

None

[fonzdm]

Hi @imgios , thanks for the hard work! I left some comments, please check'em out. Basically, two main thing to address

• Indexers json file, I do not see where it is passed to the job
• Services host reference. Where we see "servarr-" (like, servarr-radarr), we should make the servarr part variable with .Release.Namespace. This behavior is spread across yaml files and the python scripts, I pointed out something but not all of them. Please check if we can make them variable, otherwise we are forced to use the servarr name as namespace

[fonzdm]

Converted the PR to draft, hopefully we will avoid countless pipelines invocation