* Reword API Server aggregation
* Document custom resources ahead of APIService
CustomResourceDefinition is the newer, shinier and often more
appropriate resource to help with cases where APIService was not a good
fit. Switch order to mention the custom resources page first, which
introduces both APIService and CustomResourceDefinition.
* Add whats next section to PersistentVolume page
Add whats next section to PersistentVolume page
* Adding references to PersistentVolume page
Adding references to PersistentVolume page
* Use feature state for non-preempting PriorityClasses
* PreemptionPolicy requires NonPreemptingPriority feature gate
* Remove alpha qualifier from section title and decrease example heading level
* Use singular of PriorityClass in the heading
* Note CronJob timezone comes from system timezone
The timezone for the kube-controller-manager (or if that's broken apart,
the cronjob controller) determines when CronJobs are scheduled.
This reverts commit bad124b857.
* Expand details about CronJob timezone
* Consolidate words of caution about Pod network
* Tweak wording
- use tooltips
- fix a TODO hyperlink
- adopt style guidelines
* Revise prerequisites for kubeadm
* Rework page structure
- Replace some headings with anchor elements (preserving inbound links)
- Use a "discussion" section for the discussion part of the page.
- Make Feedback be a part of the What's Next section
- Skip mentioning Docker in a logging context; provide generic
signposting instead.
- Update overview
- Document limitations and fix link to HA topology
- Fixes for styling
* Redo network plugin info
* Use glossary tooltips to introduce terms
* Add Elastic metricbeat to examples of DaemonSets
The URL points to the docs related to how to configure metricbeat on k8s
* Filebeat is the next thing
* Refined unclear sentence on 3rd party dependencies
I reworded the sentence on third party dependencies a bit in order to make it more sound
* Update content/en/docs/concepts/security/overview.md
Sounds much better
Co-Authored-By: Tim Bannister <tim@scalefactory.com>
Co-authored-by: Tim Bannister <tim@scalefactory.com>
* service topology: add missing 'enabling service topology' page
Signed-off-by: Andrew Sy Kim <kiman@vmware.com>
* service topology: add common examples
Signed-off-by: Andrew Sy Kim <kiman@vmware.com>
The pod lifecycle documentation erroneously indicated `Unschedulable` as a possible `type` of pod condition. That's not true. Only four condition types exist. The `Unschedulable` value is not a type, but one of the possible reasons of the `PodScheduled` condition type.
* Added a note about build-in priority-classes
* Update content/en/docs/concepts/configuration/pod-priority-preemption.md
Co-Authored-By: Tim Bannister <tim@scalefactory.com>
Co-authored-by: Tim Bannister <tim@scalefactory.com>
* Add note for LB behaviour for cordoned nodes.
See also https://github.com/kubernetes/kubernetes/issues/65013
This is a reasonably common pitfall: `kubectl cordon <all nodes>` will also drop all LB traffic to the cluster, but this is not documented anywhere but in issues, when found it is usually already too late.
* Update with feedback
Tim Bannister