feat(ui): 411 overview dashboard core services

[BillyFigueroa]

Description

As a consumer of UDS Runtime, I would like to view a summary of core services, status and a corresponding link to view the service

Additional context

Figma Design
https://www.figma.com/design/zmKcJ9Xin7ChzyGy6RCFLe/UDS-Runtime-(UI%2FCLI)?node-id=4325-5166&m=dev

Related Issue

• Overview Dashboard: Core Services #411

Screenshots

With Services

With out Services

[TristanHoladay]

@BillyFigueroa looks good! do the displayed "names" of the services need to be changed, though, given the Figma design? (e.g. Authorization, Monitoring, Service Mesh...)

[UncleGedd]

@BillyFigueroa looks good! do the displayed "names" of the services need to be changed, though, given the Figma design? (e.g. Authorization, Monitoring, Service Mesh...)

Just chatted with @Madeline-UX about this, these names should (ideally) line up with the UDS functional layers

[UncleGedd]

Also (again just chatted with @Madeline-UX ), let's remove the "View Service >" link for now

[BillyFigueroa]

Also (again just chatted with @Madeline-UX ), let's remove the "View Service >" link for now

The links were removed in this commit. Should we ask design to remove from the mocks?

[TristanHoladay]

@BillyFigueroa sorry to add another comment 😅 but i think there's 2 issues happening. Apologies if you're already aware of them.

1. looks like the view is not re-rendering when Core services are deploying
   

2. Istio does not have a UDS Package so is not showing in Core Services for Service Mesh. Probably easiest way to get that would be to look for an istiod pod in istio-system namespace / at gateway pods in the istio-admin-gateway or istio-tentant-gateway namespaces.

[BillyFigueroa]

@BillyFigueroa sorry to add another comment 😅 but i think there's 2 issues happening. Apologies if you're already aware of them.

1. looks like the view is not re-rendering when Core services are deploying
   ![Screenshot from 2024-10-22 12-42-15](https://private-user-images.githubusercontent.com/40547442/378977900-bde8bb7a-f92f-4588-8ae9-961a9e234968.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3Mjk2MjgyNDQsIm5iZiI6MTcyOTYyNzk0NCwicGF0aCI6Ii80MDU0NzQ0Mi8zNzg5Nzc5MDAtYmRlOGJiN2EtZjkyZi00NTg4LThhZTktOTYxYTllMjM0OTY4LnBuZz9YLUFtei1BbGdvcml0aG09QVdTNC1ITUFDLVNIQTI1NiZYLUFtei1DcmVkZW50aWFsPUFLSUFWQ09EWUxTQTUzUFFLNFpBJTJGMjAyNDEwMjIlMkZ1cy1lYXN0LTElMkZzMyUyRmF3czRfcmVxdWVzdCZYLUFtei1EYXRlPTIwMjQxMDIyVDIwMTIyNFomWC1BbXotRXhwaXJlcz0zMDAmWC1BbXotU2lnbmF0dXJlPTViYzBhZWNjZTA1NDAzODkwYmUyNGFiMTM5ZTVhMDhmYjBiYzE4ZThmNjNhYTgwNDQ3YjExYjU4YTYwMzFhZTYmWC1BbXotU2lnbmVkSGVhZGVycz1ob3N0In0.OBLBGY6vpg2QWy7LpDU58r5Sks-93zxR_RTA3rVjWeg)

2. Istio does not have a UDS Package so is not showing in Core Services for Service Mesh. Probably easiest way to get that would be to look for an istiod pod in istio-system namespace / at gateway pods in the istio-admin-gateway or istio-tentant-gateway namespaces.

Hi Tristan,

1. What steps do I take to try to recreate the deploying stage?
2. I ll work on the istio changes

[TristanHoladay]

Hi Tristan,

1. What steps do I take to try to recreate the deploying stage?
2. I ll work on the istio changes

I did uds deploy k3d-core-slim-dev:latest --confirm and then, after the cluster started, deployed Runtime locally.

[decleaver]

Not sure if its just me. But pulled down the latest updates, running a newly created k3d cluster, so no core service. But now i cant select any of the views, but if I manually type in the path to view pods for example, then the view menu starts working again.....

[BillyFigueroa]

Not sure if its just me. But pulled down the latest updates, running a newly created k3d cluster, so no core service. But now i cant select any of the views, but if I manually type in the path to view pods for example, then the view menu starts working again.....

Not sure if its just me. But pulled down the latest updates, running a newly created k3d cluster, so no core service. But now i cant select any of the views, but if I manually type in the path to view pods for example, then the view menu starts working again.....

I could not re-create this. Maybe someone else can try

[BillyFigueroa]

@decleaver this might be an ignorant question because I do not understand how to handle errors in go but should this actually be considered an error when we don't have a crd? are most of these resources an array? Could we not just return an empty array in these cases?

I see this is being handled in the k8s/store.ts i.e

  if ($resources.length > 0 && typeof $resources[0].resource === 'string') {
    if ($resources[0].resource === 'crd not found') {
      this.numResources.set(0)
      return $resources
    }
  }

I remember this was causing an issue with one of the tables before.

I will add some logic in the onmessage (in cluster-overview) to handle this rather than at the component level, but maybe we should consider a different response? returning error messages to a front end like this is a bit passive. If we can make it an exception or an empty array, I think it would be cleaner

[decleaver]

@decleaver this might be an ignorant question because I do not understand how to handle errors in go but should this actually be considered an error when we don't have a crd? are most of these resources an array? Could we not just return an empty array in these cases?

I see this is being handled in the k8s/store.ts i.e

  if ($resources.length > 0 && typeof $resources[0].resource === 'string') {
    if ($resources[0].resource === 'crd not found') {
      this.numResources.set(0)
      return $resources
    }
  }

I remember this was causing an issue with one of the tables before.

I will add some logic in the onmessage (in cluster-overview) to handle this rather than at the component level, but maybe we should consider a different response? returning error messages to a front end like this is a bit passive. If we can make it an exception or an empty array, I think it would be cleaner

Yeah, so this is a little bit of a special case where if we just returned an empty array, we wouldn't be able to tell if the cluster doesn't have any of those resources or if the custom resource definition for that resource exists in the cluster, and if the CRD doesn't exist, we want to let the users know that. This could be something we revisit and find a cleaner way to handle though.

[decleaver]

paired to address last fixes