* Consider readinessGates + ready condition in diagnose
Now that readiness gates are supported, those should be considered when displaying a pod healthiness
Additionally consider the pod ready condition status. It should match the && of the containers' ready condition and readiness gates but I've actually observed that in some cases, it can be false with the containers reporting ready true.
* Update pod.go
* Update pod.go
fix lint
* Update pod.go
lint
* add tests for diagnose and readinessGateStats
* lint error
* fix#3455
* Clean up and refactor
* fix#3495#3470#3455 - fix cmd alias and filters
* fix#3535 - revert
* fix#3478 - make vulscan a bit less cpu extensive (init stab...)
* fix issue with plugin foreground exit
* fix#3466-add shared gpu on nodes
* fix#3541 - use default namespace when not specified
* update deps + release notes
* fix(logs): enhance log streaming with retry mechanism and error handling
* fix(logs): improve log tailing with enhanced retry logic and error handling
* renamed function
* Enhance log streaming with retry logic based on pod status
* Refactor Pod struct definition by moving it above the shouldStopRetrying method
* Implement exponential backoff for log streaming retries
* fix: reduce log channel buffer size to prevent drops
* Add delete functionality for Kubernetes contexts
This adds the ability to delete Kubernetes contexts using the 'd' key in the context view. The implementation includes:
- New delete command bound to 'd' key
- Confirmation dialog before deletion
- Error handling and UI refresh after deletion
* Refactor context delete dialog to use standard dialog component
- Replace custom delete confirmation with dialog.ShowConfirm
- Add read-only mode checks for rename and delete commands
- Change delete keybinding from 'd' to 'Ctrl+d'
- Remove unused deletePage constant and related functions
- Simplify delete callback logic
* Refactor context view key bindings to separate dangerous operations
* Fix linting error
* refactor: change refreshRate type to float64 for improved precision
* test: update assertions for refreshRate to use assert.InDelta for precision
* refactor: enforce minimum refresh rate and update related tests
* refactor: change refresh rate type to float32 and update related logic
* refactor: update refresh rate validation and logging in GetRefreshRate method
* refactor: update logging keys for refresh rate validation in GetRefreshRate method
Issue: when calling exec / attach on a pod with sidecar containers or ephemeral containers, those are not presented as potential targets to exec / attach / transfer.
Proposed solution: the method fetchContainers is only called with allContainers=false to find containers we can exec / attach / transfer.
Both ephemeral containers and sidecar init containers could be used in this case however with the current behavior they're not shown.
I'm proposing to change the behavior of the method to return any sidecar container and ephemeral container when called with allContainers=false since there is no other caller using that parameter set to false, so we won't break other call sites
When a CRD's GV contained k8s.io (e.g.
infrastructure.cluster.x-k8s.io/v1beta1), it was considered as standard
resource, therefore no alias was generated in the form of
"<name>.<group>", therefore the CRD list view could not open the listing
of that specific CRD.
小染
derailed
jfremy-openai
henry
Maico Timmerman
Ümüt Özalp
Alex Cantu
YE Feng
yingshanghuangqiao
Brandt Keller
Alex Trinker
Maurizio Pasquinelli
Youssef Rabie
Rudrakh Panigrahi
Vageesha B R
Max Xu
tscuite
Arriv9l
Sunggun Yu
Jason Tackaberry
D54