KEP-4368: Job Managed By; Promote to GA

Signed-off-by: Dejan Zele Pejchev <pejcev.dejan@gmail.com>

Author: dejanzele

keps/prod-readiness/sig-apps/4368.yaml

@@ -3,3 +3,5 @@ alpha:
   approver: "@wojtek-t"
 beta:
   approver: "@wojtek-t"
+stable:
+  approver: "@wojtek-t"

keps/sig-apps/4368-support-managed-by-for-batch-jobs/README.md

@@ -140,7 +140,7 @@ controller, and delegate the status synchronization to the Kueue controller.
 ### Non-Goals
 
 - passing custom parameters to the external controller
-- Introduce a new concurrency policy for CronJobs (eg. `ForbidActive` or `SoftForbid`)
+- Introduce a new concurrency policy for CronJobs (e.g. `ForbidActive` or `SoftForbid`)
 to replace a Job that is about to complete, but still has terminating pods.
 
 ## Proposal
@@ -278,7 +278,7 @@ a typo in the field value, or the cluster administrator may not install the
 custom controller (like MultiKueue) on that cluster.
 
 In such cases the user may not observe any progress by the job for a long time
-and my need to debug the Job.
+and may need to debug the Job.
 
 In order to allow for debugging of situations like this the Job controller will
 put a log line indicating the synchronization is delegated to another controller
@@ -352,9 +352,6 @@ type JobSpec struct {
 	// all characters before the first "/" must be a valid subdomain as defined
 	// by RFC 1123. All characters trailing the first "/" must be valid HTTP Path
 	// characters as defined by RFC 3986. The value cannot exceed 64 characters.
-	//
-	// This field is alpha-level. The job controller accepts setting the field
-	// when the feature gate JobManagedBy is enabled (disabled by default).
 	// +optional
 	ManagedBy *string
 }
@@ -493,9 +490,9 @@ The following scenarios are covered:
   - it does not add the Suspended condition for a Job with `.spec.suspend=true` ([link](https://github.com/kubernetes/kubernetes/blob/856475e5fffe3d99c71606d6024f5ed93e37eebc/test/integration/job/job_test.go#L2059)).
 - the Job controller reconciles jobs with custom "managedBy" field when the feature gate is disabled ([link](https://github.com/kubernetes/kubernetes/blob/856475e5fffe3d99c71606d6024f5ed93e37eebc/test/integration/job/job_test.go#L2030))
 - the Job controller handles correctly re-enablement of the feature gate [link](https://github.com/kubernetes/kubernetes/blob/169a952720ebd75fcbcb4f3f5cc64e82fdd3ec45/test/integration/job/job_test.go#L1691)
-- the `job_by_external_controller_total` metric is incremented when a new Job with custom "managedBy" is created ([link](https://github.com/kubernetes/kubernetes/blob/856475e5fffe3d99c71606d6024f5ed93e37eebc/test/integration/job/job_test.go#L2044-L2058))
-- the `job_by_external_controller_total` metric is not incremented for a new Job without "managedBy" or with default value ([link](https://github.com/kubernetes/kubernetes/blob/856475e5fffe3d99c71606d6024f5ed93e37eebc/test/integration/job/job_test.go#L2000-L2029))
-- the `job_by_external_controller_total` metric is not incremented for Job updates (regardless of the "managedBy") (tested indirectly as [here](https://github.com/kubernetes/kubernetes/blob/856475e5fffe3d99c71606d6024f5ed93e37eebc/test/integration/job/job_test.go#L2000-L2029) the Job controller updates the Job status)
+- the `jobs_by_external_controller_total` metric is incremented when a new Job with custom "managedBy" is created ([link](https://github.com/kubernetes/kubernetes/blob/856475e5fffe3d99c71606d6024f5ed93e37eebc/test/integration/job/job_test.go#L2044-L2058))
+- the `jobs_by_external_controller_total` metric is not incremented for a new Job without "managedBy" or with default value ([link](https://github.com/kubernetes/kubernetes/blob/856475e5fffe3d99c71606d6024f5ed93e37eebc/test/integration/job/job_test.go#L2000-L2029))
+- the `jobs_by_external_controller_total` metric is not incremented for Job updates (regardless of the "managedBy") (tested indirectly as [here](https://github.com/kubernetes/kubernetes/blob/856475e5fffe3d99c71606d6024f5ed93e37eebc/test/integration/job/job_test.go#L2000-L2029) the Job controller updates the Job status)
 
 The following scenarios related to [Terminating pods and terminal Job conditions](#terminating-pods-and-terminal-job-conditions) are covered:
 - `Failed` or `Complete` conditions are not added while there are still terminating pods ([link](https://github.com/kubernetes/kubernetes/blob/856475e5fffe3d99c71606d6024f5ed93e37eebc/test/integration/job/job_test.go#L1183))
@@ -523,7 +520,7 @@ We propose a single e2e test for the following scenario:
   API fields affected by the new validation rules
 - make CronJob more resilient by checking the Job condition is `Complete` when using `CompletionTime` (see [here](#custom-controllers-not-compatible-with-api-assumptions-by-cronjob))
 - The feature flag disabled by default
-- implement the `job_by_external_controller_total` metric
+- implement the `jobs_by_external_controller_total` metric
 
 Second Alpha (1.31):
 - preparatory fix to address all known inconsistencies between validation and the
@@ -548,7 +545,7 @@ Second Alpha (1.31):
 - Address reviews and bug reports from Beta users
 - Re-evaluate the ideas of improving debuggability (like [extended `kubectl`](#debuggability), [dedicated condition](#condition-to-indicated-job-is-skipped), or [events](#event-indicating-the-job-is-skipped))
 - Re-evaluate the need to skip reconciliation in the event handlers to optimize performance
-- Assess the fragmentation of the ecosystem. Look for other implementations of a job controller and asses their conformance with k8s.
+- Assess the fragmentation of the ecosystem. Look for other implementations of a job controller and assess their conformance with k8s.
 - Lock the feature gate
 
 #### Deprecation
@@ -752,7 +749,7 @@ being flipped between two owners.
 The feature is opt-in so in case of such problems the custom "managedBy" field
 should not be used.
 
-Also, an admin could check if the value of the `job_by_external_controller_total`
+Also, an admin could check if the value of the `jobs_by_external_controller_total`
 matches the expectations. For example, if the value of the metric does not increase
 when new jobs are being added with a custom "managedBy" field, it might be
 indicative that the feature is not working correctly.
@@ -831,7 +828,7 @@ checking if there are objects with field X set) may be a last resort. Avoid
 logs or events for this purpose.
 -->
 
-Check the `job_by_external_controller_total` metric. If the value is non-zero
+Check the `jobs_by_external_controller_total` metric. If the value is non-zero
 for a field, it means there were Jobs using the custom controller created, so
 the feature is in use.
 
@@ -885,22 +882,24 @@ Pick one more of these and delete the rest.
 
 - [x] Metrics
   - Metric name:
-    - `job_by_external_controller_total` (new), with the `controller_name` label
+    - `jobs_by_external_controller_total` (new), with the `controller_name` label
 corresponding to the custom value of the "managedBy" field. The metric is
 incremented by the built-in Job controller on each ADDED Job event,
 corresponding to a Job with custom value of the "managedBy" field.
 This metric can be helpful to determine the health of a job and its controller
 in combination with already existing metrics (see below).
+      - Components exposing the metric: kube-controller-manager
     - `apiserver_request_total[code=409, resource=job, group=batch]` (existing):
-substantial increase of this metric, when additionally `job_by_external_controller_total>0`
+substantial increase of this metric, when additionally `jobs_by_external_controller_total>0`
 may be indicative of two controllers stepping onto each-other causing
 conflicts (see [here](#two-controllers-running-at-the-same-time-on-old-version)).
+      - Components exposing the metric: kube-apiserver
     - `kube_cronjob_status_active` (existing), substantial increase of this
 metric, may suggest that there are accumulating non-progressing jobs controlled
-by `CronJob`. If additionally `job_by_external_controller_total>0` it may suggest
+by `CronJob`. If additionally `jobs_by_external_controller_total>0` it may suggest
 that the Jobs are getting stuck due to not being synchronized by the custom
 controller.
-  - Components exposing the metric: kube-apiserver
+      - Components exposing the metric: kube-apiserver
 
 ###### Are there any missing metrics that would be useful to have to improve observability of this feature?
 
@@ -1080,7 +1079,7 @@ N/A.
 
 ## Implementation History
 
-- 2023-12.20 - First version of the KEP
+- 2023-12-20 - First version of the KEP
 - 2024-03-05 - Merged implementation PR [Support for the Job managedBy field (alpha)](https://github.com/kubernetes/kubernetes/pull/123273)
 - 2024-03-07 - Merged [Update Job conformance test for job status updates](https://github.com/kubernetes/kubernetes/pull/123751)
 - 2024-03-08 - Merged [Follow up fix to the job status update test](https://github.com/kubernetes/kubernetes/pull/123815)
@@ -1090,6 +1089,7 @@ N/A.
 - 2024-06-21 - Merged [Update the count of ready pods when deleting pods](https://github.com/kubernetes/kubernetes/pull/125546)
 - 2024-07-12 - Merged [Delay setting terminal Job conditions until all pods are terminal](https://github.com/kubernetes/kubernetes/pull/125510)
 - 2024-07-30 - Merged [Update the docs for JobManagedBy and JobPodReplacementPolicy related to pod termination](https://github.com/kubernetes/website/pull/46808)
+- 2024-10-17 - Merged [Graduate JobManagedBy to Beta in 1.32](https://github.com/kubernetes/kubernetes/pull/127402)
 
 <!--
 Major milestones in the lifecycle of a KEP should be tracked in this section.
@@ -1369,7 +1369,7 @@ It makes them not that useful to debug situations when the Job didn't make
 progress for long time. So, they would not give a reliable signal for debugging
 based on playbooks.
 
-Renewing the even on every Job update seems excessive from the performance
+Renewing the event on every Job update seems excessive from the performance
 perspective.
 
 ## Infrastructure Needed (Optional)

keps/sig-apps/4368-support-managed-by-for-batch-jobs/kep.yaml

@@ -3,7 +3,7 @@ kep-number: 4368
 authors:
   - "@mimowo"
 owning-sig: sig-apps
-status: implementable
+status: implemented
 creation-date: 2023-12-20
 reviewers:
   - "@alculquicondor"
@@ -18,17 +18,18 @@ see-also:
   - "https://github.com/kubernetes/enhancements/pull/4073" # closed PR
 
 # The target maturity stage in the current dev cycle for this KEP.
-stage: beta
+stage: stable
 
 # The most recent milestone for which work toward delivery of this KEP has been
 # done. This can be the current (upcoming) milestone, if it is being actively
 # worked on.
-latest-milestone: "v1.32"
+latest-milestone: "v1.35"
 
 # The milestone at which this feature was, or is targeted to be, at each stage.
 milestone:
   alpha: "v1.30"
   beta: "v1.32"
+  stable: "v1.35"
 
 # The following PRR answers are required at alpha release
 # List the feature gate name and the components for which it must be enabled
@@ -40,4 +41,4 @@ feature-gates:
 disable-supported: true
 
 metrics:
-- job_by_external_controller_count
+- jobs_by_external_controller_total