explain Available condition behavior in deployment concepts #39892
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
k8s-ci-robot
added
cncf-cla: yes
| `type: Available` with `status: "True"` means that your Deployment is healthy and has a minimum availability (`.status.availableReplicas` is at least equal to `.spec.replicas - maxUnavailable`). | ||
| `maxUnavailable` is computed according to the deployment strategy: | ||
| - For the `RollingUpdate` strategy, it is a resolved value of `.spec.strategy.rollingUpdate.maxUnavailable`. | ||
| - For the `Recreate` strategy, it is always equal to 100% of the replicas that we allow to be down, so that new rollouts can proceed. |
There was a problem hiding this comment.
This counts with a change proposed in kubernetes/kubernetes#113228. Currently it is 0.
✅ Pull request preview available for checkingBuilt without sensitive environment variables
To edit notification comments on pull requests, go to your Netlify site settings. |
| The `Available` condition is designed to support rollouts first. If `maxUnavailable` is equal to `.spec.replicas`, | ||
| then this condition will never become `status: "False"`, even if no pods are running. | ||
| In this case you should check `.status.availableReplicas` to be sure that at least 1 pod is present. | ||
| It is recommended to have the number of `.spec.replicas` higher than `maxUnavailable`, to have always some pods available during a rollout. |
There was a problem hiding this comment.
Thanks. I suppose we will wait with this PR until we are all satisfied with the changes in kubernetes/kubernetes#113228
|
@dims I think kubernetes/kubernetes#113228 if we want to change the |
|
Or we could publish the documentation without mentioning the |
|
I would prefer to get the update to the strategy - I think we made the case in 113228 that we should be consistent and the current behavior provides no value for end users on recreate (they can only be wrong). |
|
👋 @atiratree . |
|
/hold |
k8s-ci-robot
added
the
do-not-merge/hold
|
Hello @atiratree . |
|
@smarterclayton would you have time to revisit kubernetes/kubernetes#113228? Or should I overtake it, or try to wrap this one first? |
|
Hello @atiratree , please could you advise where we are with these changes? |
1 similar comment
|
Hello @atiratree , please could you advise where we are with these changes? |
|
There hasn't been an update in a long while. I will close this PR. /close |
|
@sftim: Closed this PR. In response to this:
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. |
We need to come to a consensus how to define an Available condition. This has been already discussed in these places
cc @smarterclayton @soltysh @janetkuo @alculquicondor @ravisantoshgudimetla