Re: [kubernetes/kubernetes] kubectl get pods should report last restart time? (#54491)
grodrigues3
@kubernetes/sig-cli-feature-requests
—
You are receiving this because you are on a team that was mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
zhengjiajin
i thank we can add a column LASTSTARTED. alse it should be close to RESTARTS column
zhengjiajin
/assign
Stewart-YU
I think add a column LASTSTARTED need, keeping RESTARTS column also need. LASTSTARTED reflect short-term pod's status, RESTARTS reflect long-term pod's status recently. @zjj2wry
zhengjiajin
when my pod do not have lastState:(my pod not running all the time). should the column show?
David Oppenheimer
I have a REALLY dumb question. What causes a pod to restart? (As opposed to being deleted.)
VictorBHub
@zjj2wry 'lastState:' contains info about the start and end of the last dead pod.
the original suggestion, I think, was to show the current pod ready since time (the 'lastTransitionTime' under status.conditions on the element where "type": "Ready")
fejta-bot
Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.
If this issue is safe to close now please do so with /close.
Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale
Eric Sniff
@davidopp in our stack the main causes are OOMs and segfaults.
fejta-bot
Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.
If this issue is safe to close now please do so with /close.
Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten
/remove-lifecycle stale
fejta-bot
Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.
Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close
k8s-ci-robot
Closed #54491.
etibroo
/assign
What happened after this? Is there a plan to include this?
HerrTamm
Also interested in the smae feature, it especially helps to debug services in general namespaces (e.g. kube-system), where multiple services and a lot of pods are running.
—
You are receiving this because you are on a team that was mentioned.
Reply to this email directly, view it on GitHub, or unsubscribe.
rohitkhatana
This feature should be there. It helps in identifying issue in much better/easy way. For example if we are facing issues during in a time window, in the that time window it is very relevant information to see.
antoniocascais
Totally agree with @rohitkhatana !
Pieter
I have a REALLY dumb question. What causes a pod to restart? (In-place on the same node. As opposed to being deleted.)
I think the container exiting on it's own, e.g. app detecting a missing value at startup or faulty db connection and then electing to exit.
Maciej Szulik
/reopen
/lifecycle frozen
Maciej Szulik
/assign
Kubernetes Prow Robot
@soltysh: Reopened this issue.
In response to this:
/reopen
/lifecycle frozen
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.
Kubernetes Prow Robot
Reopened #54491.
desertfury
/assign
Maciej Szulik
For anyone willing to work on this, the last restart time should be an approximation from all the statuses of all containers and initcontainers. I will want to see extensive unit tests for this particular case.
Jordan Jacobelli
Hey, I would like to work on this if possible. That would be my first contribution to Kubernetes so I may need help at some point
Maciej Szulik
@Ethyling go ahead, feel free to ping me on slack, but it might take a bit for me to reply 😉 due to upcoming code freeze for 1.21.
Jordan Jacobelli
@Ethyling go ahead, feel free to ping me on slack, but it might take a bit for me to reply wink due to upcoming code freeze for 1.21.
Hey @soltysh , thank you!
I currently have a first version:
$ kubectl get pods
NAME READY STATUS RESTARTS LAST RESTART AGE
nginx-deployment-66b6c48dd5-sb297 1/1 Running 1 9m30s 12m
nginx-deployment-79fbfbd7c8-hj8z5 0/1 Init:CrashLoopBackOff 5 54s 4m40s
I'm using the LastTerminationState attribute of the container status. I don't know if it's the right attribute to use. I will clean my code and open a PR. I think it will be easier to discuss
Jordan Jacobelli
Hey @soltysh, I opened the PR, if you have some time to check it