feat: Add metric to expose Applications conditions

[foyerunix]

Hello,

Closes #13096

Implement a new metric exposing Applications conditions. This is particularly useful for SRE teams to be able to setup alerts on issues that aren't displayed via "health_status" and "sync_status" in the metric "argocd_app_info".

Our use case is similar to #13096. We want to ensure flawless reinstallation of our Kubernetes clusters. To do this we need to take care to not have some resources that for some reason aren't managed by ArgoCD but that end up being necessary for our workloads to run properly.

Best Regards.

[bunnyshell]

❌ Preview Environment deleted from Bunnyshell

Available commands (reply to this comment):

• 🚀 /bns:deploy to deploy the environment

[bunnyshell]

❌ Preview Environment deleted from Bunnyshell

Available commands (reply to this comment):

• 🚀 /bns:deploy to deploy the environment

[codecov]

Codecov Report

Attention: Patch coverage is 96.66667% with 1 line in your changes missing coverage. Please review.

Project coverage is 55.90%. Comparing base (5834175) to head (0beb22d).
Report is 40 commits behind head on master.

Files with missing lines	Patch %	Lines
cmd/argocd/commands/admin/app.go	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #19438      +/-   ##
==========================================
+ Coverage   55.83%   55.90%   +0.07%     
==========================================
  Files         320      320              
  Lines       44041    44061      +20     
==========================================
+ Hits        24589    24634      +45     
+ Misses      16896    16871      -25     
  Partials     2556     2556              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[ilia-medvedev-codefresh]

@foyerunix I think the current implementation does not take into account that a condition can appear more than just one time. For example an application can have multiple excluded resources and those would all be appended to the application status

argo-cd/controller/state.go

Line 509 in 3feab7a

conditions = append(conditions, v1alpha1.ApplicationCondition{

.
You should instead count the appearances of each condition and increase the gauge value accordingly.

Another concern I have about this is scale. An application can potentially have 9 different condition types and more can be added in the future. This change would generate a gauge per each existing status. Perhaps we should add a parameter with a whitelist of statuses that we want to collect, similar to how it's done with application labels.

[foyerunix]

Hello @ilia-medvedev-codefresh,

I took into account your propositions and updated the pull request.

Best Regards.

[pasha-codefresh]

LGTM, thank you