castai / k8s-agent Goto Github PK

Hello, I would like to use the cost analysis feature with the dashboard but wish to self-host cast.ai. Is that possible? Or do I need to send data from my cluster to your services?

We use some custom deployment CRD’s (argo rollouts - Rollout) to drive our workloads - is it possible to include those into cost reporting? They answer to all the same API’s as Deployment does - they’re just an extension.

Here's a spec: https://argoproj.github.io/argo-rollouts/features/specification/

Minimal example:

apiVersion: argoproj.io/v1alpha1
kind: Rollout
metadata:
  name: rollouts-demo
spec:
  replicas: 5
  strategy:
    canary:
      steps:
      - setWeight: 20
      - pause: {}
      - setWeight: 40
      - pause: {duration: 10}
      - setWeight: 60
      - pause: {duration: 10}
      - setWeight: 80
      - pause: {duration: 10}
  revisionHistoryLimit: 2
  selector:
    matchLabels:
      app: rollouts-demo
  template:
    metadata:
      labels:
        app: rollouts-demo
    spec:
      containers:
      - name: rollouts-demo
        image: argoproj/rollouts-demo:blue
        ports:
        - name: http
          containerPort: 8080
          protocol: TCP
        resources:
          requests:
            memory: 32Mi
            cpu: 5m

Hello.

I have a "problem" that would be some workloads that have their CPU and Memory consumption data blank, so we don't have a cost calculation, there are several workloads in several different clusters. Our clusters are Azure Kubernetes Services.

Initially I imagined that the cause was due to the low consumption of resources, some with CPU in 0.01 of use, but the memory always has a consumption that should already be relevant (20Mi, others with 60Mi, etc) for the cost calculations.

I would like some guidance on what to evaluate to find out the reason for the cost analysis at 0.

Thank you very much!

Looks like the only way to install cast.ai is manually running some shell command which includes kubectl apply. This goes against best practices and common sense in production environments where one would manage all infrastructure, including Kubernetes resources as code.
Since Helm is an industry standard, I think it would make sense to provide a Helm chart and relevant documentation on non-shell command installation method.