Show pod status in Workloads / Pods view

[alexbrand]

Kubectl seems to do some magic to show a "STATUS" of a given pod:

NAME                       READY   STATUS              RESTARTS   AGE
busybox-644dbf5f4b-2sjfs   1/1     Terminating         0          42m
busybox-644dbf5f4b-l5lzn   0/1     ContainerCreating   0          31m

I was somewhat surprised when I used Octant to look at the same set of pods. This is what Octant shows:

Notice how kubectl says the first pod is Terminating and the second is ContainerCreating, whereas Octant shows Running and Pending.

I understand that Octant is showing the pod's Phase and not the Status (I am actually not sure how kubectl computes this), but it was super confusing at first.

I am wondering if we should follow kubectl here and show Status instead of Phase, or perhaps add a new column that shows Status.

[wwitzel3]

We are planning to refactor the way we do object status and this will be addressed at that time. Though it still may not map exactly to what you see in kubectl output, it will be improved to provide a better level of detail over just pending / running.

[wwitzel3]

Both status and phase should be displayed in Octant. Example YAML below of a Pod status / phase.

  containerStatuses:
  - image: quay.io/covalent/demo-elasticsearch:v1.2
    imageID: ""
    lastState: {}
    name: elasticsearch
    ready: false
    restartCount: 0
    started: false
    state:
      waiting:
        message: Back-off pulling image "quay.io/covalent/demo-elasticsearch:v1.2"
        reason: ImagePullBackOff
  hostIP: 10.128.15.203
  phase: Pending
  podIP: 10.48.7.15
  podIPs:
  - ip: 10.48.7.15
  qosClass: BestEffort
  startTime: "2021-07-14T17:49:16Z"

[ftovaro]

This one is pretty clear of what should be done, I'd like to take this one and work on it

[ftovaro]

I've been working on this one, the only problem I'm having is with the Terminated state.

When we are in a Pending phase like when we are creating a Pod, I can extract a Reason message from the Pod struct that says "ContainerCreating" as we need.

➜  ~ kubectl get all
NAME             READY   STATUS              RESTARTS   AGE
pod/client-pod   0/1     ContainerCreating   0          1s

But, when I get a Terminated state this is the output from kubectl get all:

➜  ~ kubectl get all
NAME             READY   STATUS        RESTARTS   AGE
pod/client-pod   0/1     Terminating   0          65s

And the Reason message from the Pod struct says Completed

Also, is visible in the image that with a Terminated state the phase is still on Running

This is the only inconsistency I'm having to finish the issue. Here is the draft PR #2675

[wwitzel3]

@ftovaro and I met up, we are going to us pod.DeletionTimestamp != nil && lastStatus.State.Terminated == nil to infer the Terminating status.