## What this PR does
Fixes the "Pod serving" (EndpointSlice) table on the service details
page which displayed "Raw:" prefixes and "Invalid Date" values. The root
cause was a missing `CustomColumnsOverride` for the
`factory-kube-service-details-endpointslice` customization ID referenced
by the `EnrichedTable` component.
Adds column definitions: Pod (`.targetRef.name`), Addresses
(`.addresses`), Ready (`.conditions.ready`), Node (`.nodeName`).
<img width="1113" height="260" alt="Pod serving table"
src="https://github.com/user-attachments/assets/5bf92080-feee-44b3-942f-a019ba388427"
/>
### Release note
```release-note
[dashboard] Fix EndpointSlice rendering in service details — Pod serving table now correctly displays pod name, addresses, readiness, and node.
```
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **New Features**
* Added new columns to the service endpoints view in the dashboard,
displaying Pod name, IP addresses, Ready status, and assigned Node
information for better visibility into pod serving details.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
## What this PR does
Adds missing `baseFactoriesMapping` entries for backup resources (plans,
backupjobs, backups) to the static Navigation resource.
Backup resources are not ApplicationDefinitions, so `ensureNavigation()`
never creates their navigation mappings. Without these mappings, the
OpenUI frontend cannot resolve the `{cluster}` context for backup pages,
producing broken sidebar links with an empty cluster segment (e.g.
`/openapi-ui//tenant-root/api-table/backups.cozystack.io/v1alpha1/plans`
instead of `/openapi-ui/default/tenant-root/...`).
### Release note
```release-note
[dashboard] Fix broken backup menu links by adding missing baseFactoriesMapping entries for backup resources
```
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **New Features**
* Backup resources are now integrated into dashboard navigation,
providing quick access to backup plans, backup job monitoring, and
detailed backup information for streamlined operations management.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
The service details page showed "Raw:" and "Invalid Date" in the Pod
serving table because the EnrichedTable referenced customizationId
"factory-kube-service-details-endpointslice" which had no corresponding
CustomColumnsOverride. Add column definitions for Pod, Addresses, Ready,
and Node fields.
Assisted-By: Claude AI
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
Import SchedulingClassAnnotation, SchedulingClassLabel, SchedulerName,
and GVR components from the cozystack-scheduler/pkg/apis sub-module
instead of defining them locally. This ensures the webhook and scheduler
stay in sync on label/annotation keys.
Also standardize the namespace label from scheduling.cozystack.io/class
to scheduler.cozystack.io/scheduling-class for consistency with the
scheduler, and resolve scheduling class from the owner Application CR
(via a new SchedulingClass() stub method) before falling back to the
namespace label.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Signed-off-by: Timofei Larkin <lllamnyp@gmail.com>
Backup resources (plans, backupjobs, backups) are not
ApplicationDefinitions, so ensureNavigation() never adds their
baseFactoriesMapping entries. Without these mappings the OpenUI
frontend cannot resolve the {cluster} context for backup pages,
producing broken sidebar links with an empty cluster segment
(e.g. /openapi-ui//tenant-root/...).
Add the three missing entries to the static Navigation resource.
Assisted-By: Claude AI
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
Before injecting schedulerName and annotation, the webhook now
checks that the referenced SchedulingClass CR actually exists.
If the CRD is not installed or the CR is missing, injection is
skipped so that pods are not stuck Pending on a non-existent
scheduler.
Assisted-By: Claude AI
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
When a namespace carries the scheduling.cozystack.io/class label,
the webhook injects schedulerName=cozystack-scheduler and the
scheduler.cozystack.io/scheduling-class annotation into every Pod.
This covers all workloads in the tenant namespace regardless of
whether the operator CRD supports schedulerName natively.
Assisted-By: Claude AI
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
Add an API-backed listInput dropdown for the schedulingClass field
in the Tenant creation form. The dropdown lists available
SchedulingClass CRs from cozystack.io/v1alpha1.
Assisted-By: Claude AI
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
## Summary
- Sidebar menu was showing all resources regardless of their
MarketplacePanel `hidden` state
- Fetch MarketplacePanels during sidebar reconciliation and skip
resources where `hidden=true`
- Hiding a resource from the marketplace now also removes it from the
sidebar navigation
## Test plan
- [ ] Set `hidden: true` on a MarketplacePanel (e.g. qdrant)
- [ ] Trigger controller reconciliation
- [ ] Verify the resource is removed from the sidebar menu
- [ ] Set `hidden: false` and verify the resource reappears in the
sidebar
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **New Features**
* Sidebar can now hide resources based on MarketplacePanel configuration
parsed from panel definitions.
* Hidden resources are filtered early when assembling sidebar
categories, preventing them from contributing to menu items.
* Listing failures are non-fatal: if configuration fetch fails, no
hiding is applied and the dashboard remains functional.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
## Summary
- Fix "Disabling features from menu and marketplace is not working" by
preserving user-set `disabled` and `hidden` values during controller
reconciliation
- The controller was hardcoding `disabled=false` and `hidden=false` on
every reconcile loop, overwriting any changes made through the dashboard
UI
## Test plan
- [ ] Disable a service from the dashboard marketplace panel
- [ ] Verify the service stays disabled after controller reconciliation
- [ ] Hide a service from the dashboard menu
- [ ] Verify the service stays hidden after controller reconciliation
- [ ] Create a new ApplicationDefinition and verify its MarketplacePanel
defaults to disabled=false, hidden=false
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **Bug Fixes**
* Fixed an issue where user-configured "disabled" and "hidden" settings
in the marketplace panel could be reset during updates. These
preferences are now preserved when the panel is created or updated, and
the system avoids applying unnecessary configuration changes when values
haven't actually changed.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
The external-ips factory used incorrect EnrichedTable properties causing
empty rows in the dashboard. Replace `clusterNamePartOfUrl` with
`cluster` and change `pathToItems` from array to dot-path string format
to match the convention used by all other working EnrichedTable instances.
Signed-off-by: IvanHunters <xorokhotnikov@gmail.com>
The controller was hardcoding disabled=false and hidden=false on every
reconciliation, overwriting any user changes made through the dashboard
UI. Move spec building inside the CreateOrUpdate mutate function to read
and preserve current disabled/hidden values from the existing resource.
Signed-off-by: IvanHunters <xorokhotnikov@gmail.com>
The sidebar was generated independently from MarketplacePanels, always
showing all resources regardless of their hidden state. Fetch
MarketplacePanels during sidebar reconciliation and skip resources
where hidden=true, so hiding a resource from the marketplace also
removes it from the sidebar navigation.
Signed-off-by: IvanHunters <xorokhotnikov@gmail.com>
## What this PR does
Replaces the plain text input for `storageClass` fields with an
API-backed dropdown listing available StorageClasses from the cluster.
Follows the same pattern as the `instanceType` dropdown for VMInstance.
Affected applications:
- **Top-level `spec.storageClass`**: ClickHouse, Harbor, HTTPCache,
Kubernetes, MariaDB, MongoDB, NATS, OpenBAO, Postgres, Qdrant,
RabbitMQ, Redis, VMDisk
- **Nested `spec.storage.storageClass`**: FoundationDB
- **Nested `spec.kafka.storageClass` / `spec.zookeeper.storageClass`**:
Kafka
### Release note
```release-note
[dashboard] storageClass fields in stateful app forms now render as a
dropdown populated with available StorageClasses from the cluster,
instead of a free-text input.
```
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
## Release Notes
* **New Features**
* Storage class selection dropdowns now available in configuration forms
for multiple database, messaging, and storage services.
* **Tests**
* Added comprehensive test coverage for storage class configuration
handling.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
## What this PR does
Restores `stock-instance-api-form`, `stock-instance-api-table`,
`stock-instance-builtin-form`, and `stock-instance-builtin-table`
sidebar
resources that were removed in #2106, and adds them back to the orphan
cleanup allowlist.
PR #2106 removed these sidebars to fix broken URLs on the main page
before
namespace selection (`default//api-table/...`). However,
`stock-instance-*`
sidebars are required by the frontend for namespace-level
api-table/api-form
pages. Without them and with `CUSTOMIZATION_SIDEBAR_FALLBACK_ID=""`, the
frontend cannot find a sidebar for pages like Backup Plans and renders
an
empty page where no interaction is possible.
The broken-URL bug is already fully fixed by
`CUSTOMIZATION_SIDEBAR_FALLBACK_ID=""`
in `web.yaml`. Re-adding `stock-instance-*` does not reintroduce it,
since
these sidebars are only shown when the user is on a namespace-level page
where the `{namespace}` placeholder is filled.
### Release note
```release-note
[dashboard] fix empty page on Backup Plans and other namespace-level api-table pages
```
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **New Features**
* Added four new dashboard sidebar resources for stock instances: API
form, API table, built-in form, and built-in table views. These enable
expanded dashboard customization options for managing stock instance
configurations and data.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
PR #2106 removed stock-instance-* sidebar resources to fix broken URLs
on the main page before namespace selection. However, these sidebars are
required for rendering namespace-level pages (api-table, api-form, etc.)
such as the Backup Plans page.
Without stock-instance-api-table, the frontend cannot find the sidebar
for namespace-scoped api-table pages and renders an empty page instead.
The original bug (broken URLs with empty namespace placeholder) is already
fixed by CUSTOMIZATION_SIDEBAR_FALLBACK_ID="" in web.yaml, so re-adding
stock-instance-* sidebars does not reintroduce it.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
Update openapi-k8s-toolkit commit to d6b9e4ad (release/1.4.0) which
includes the FormListInput layout refactor, making formlistinput-value-binding.diff
obsolete.
Set allowEmpty: true on the VMInstance instanceType field so users can
explicitly clear the selection and override the default instance type.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
Replace plain text input with an API-backed listInput dropdown for
storageClass fields across all applications that expose them.
The dropdown fetches available StorageClasses from the cluster via
/api/clusters/{cluster}/k8s/apis/storage.k8s.io/v1/storageclasses,
following the same pattern as the instanceType dropdown for VMInstance.
Top-level spec.storageClass: ClickHouse, Harbor, HTTPCache, Kubernetes,
MariaDB, MongoDB, NATS, OpenBAO, Postgres, Qdrant, RabbitMQ, Redis, VMDisk.
Nested paths: FoundationDB (spec.storage.storageClass),
Kafka (spec.kafka.storageClass and spec.zookeeper.storageClass).
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
Remove stock-instance-* sidebars that were populated with namespace-scoped
menu items, causing the sidebar to incorrectly appear on cluster-level pages.
Clear the sidebar fallback ID so the frontend gracefully renders no sidebar
when no matching sidebar resource exists.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
## What this PR does
Hide Ingresses, Services, and Secrets tabs in the dashboard when the
ApplicationDefinition has no resource selectors (Include/Exclude) for
the corresponding type. Previously all tabs were always hardcoded as
visible.
### Release note
```release-note
[dashboard] Hide Ingresses/Services/Secrets tabs when ApplicationDefinition has no resource selectors defined
```
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **Bug Fixes**
* Dashboard tab visibility refined: Workloads tab remains always
visible; Ingresses, Services, and Secrets tabs now appear only when
corresponding resource selectors are configured, reducing clutter and
showing relevant tabs based on configured resources.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
Only show Ingresses/Services/Secrets tabs when Include selectors are
defined. Exclude selectors alone don't make resources visible as tenant
resources, so tabs would be empty.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
Show these tabs only when the ApplicationDefinition has non-empty
Include or Exclude resource selectors for the corresponding type.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
Add API-backed listInput dropdown for disks[].name in VMInstance form,
listing available VMDisk resources from the same namespace.
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
Override spec.instanceType field with listInput type in schema so the
dashboard renders it as a select dropdown populated from
VirtualMachineClusterInstancetype resources. Default value is read
dynamically from the ApplicationDefinition's OpenAPI schema.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
CollectCRDNames now requires both apiVersion "apiextensions.k8s.io/v1"
and kind "CustomResourceDefinition", consistent with the validation in
crdinstall.Install.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
- Remove duplicate "Starting controller manager" log before install
phases, keep only the one before mgr.Start()
- Rename misleading test "document without kind returns error" to
"decoder rejects document without kind" to match actual behavior
- Document Helm uninstall CRD behavior in deployment template comment
- Use --health-probe-bind-address=0 consistently with metrics-bind
- Exclude all dotfiles in verify-crds diff, not just .gitattributes
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
The workload reconciler was refactored to use the label
workloads.cozystack.io/monitor but the test still used the old
workloadmonitor.cozystack.io/name label, causing the reconciler to
delete the workload instead of keeping it.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
- Check both apiVersion and kind when validating embedded CRD manifests
to prevent applying objects with wrong API group
- Move ctrl.SetupSignalHandler() before install phases so CRD and Flux
installs respect SIGTERM instead of blocking for up to 2 minutes
- Replace custom contains/searchString helpers with strings.Contains
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
Validate that all parsed objects are CustomResourceDefinition before
applying with force server-side apply. This prevents accidental
application of arbitrary resources if a non-CRD file is placed in
the manifests directory.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
Move duplicated YAML parsing (ReadYAMLObjects, ParseManifestFile) and
CRD readiness check (WaitForCRDsEstablished, CollectCRDNames) into a
shared internal/manifestutil package. Both crdinstall and fluxinstall
now import from manifestutil instead of maintaining identical copies.
Replace fluxinstall's time.Sleep(2s) after CRD apply with proper
WaitForCRDsEstablished polling, matching the crdinstall behavior.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
- Use explicit apiextensions.k8s.io/v1 CRD GVK in waitForCRDsEstablished
instead of fragile objects[0].GroupVersionKind()
- Add TestInstall_crdNotEstablished for context timeout path
- Add --recursive to diff in verify-crds Makefile target
- Document why both crds/ and --install-crds exist in deployment template
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
- Wait for CRDs to have Established condition after server-side apply,
instead of returning immediately
- Add TestInstall with fake client and interceptor to simulate CRD
establishment
- Add TestInstall_noManifests and TestInstall_writeManifestsFails for
error paths
- Fix fluxinstall/manifests.embed.go: use filepath.Join for OS paths
and restrict permissions from 0666 to 0600 (same fix as crdinstall)
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
- Use filepath.Join instead of path.Join for OS file paths
- Restrict extracted manifest permissions from 0666 to 0600
- Add unit tests for readYAMLObjects, parseManifests, and
WriteEmbeddedManifests including permission verification
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
Embed Package and PackageSource CRDs in the operator binary using Go
embed, following the same pattern as --install-flux. The operator applies
CRDs at startup using server-side apply, ensuring they are updated on
every operator restart/upgrade.
This addresses the CRD lifecycle concern: Helm crds/ directory handles
initial install, while the operator manages updates on subsequent
deployments.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Aleksei Sviridkin <f@lex.la>
- Upgrade CRDs in CozyStack dashboard controller
- Add Ingress proxy timeouts for WebSocket to work without terminations
- Add CFOMapping custom resource
Signed-off-by: Kirill Ilin <stitch14@yandex.ru>
## What this PR does
Fixes a race condition where multiple Packages sharing the same
namespace could overwrite each other's PodSecurity Admission labels via
Server-Side Apply.
When two PackageSources deploy to the same namespace (e.g.
`cozystack.linstor` and `cozystack.linstor-scheduler` both use
`cozy-linstor`), each Package only knew about its own components. If a
Package without `privileged: true` reconciled the namespace, SSA would
remove the `pod-security.kubernetes.io/enforce=privileged` label set by
the other Package.
On Talos clusters (which default to `baseline` PodSecurity enforcement),
this caused privileged pods like LINSTOR satellites to be rejected.
The fix makes each Package check ALL PackageSources and their active
Packages when reconciling a namespace. A namespace is set to
`privileged` if ANY Package has a component with `privileged: true` in
it. This ensures a consistent, holistic decision regardless of
reconciliation order.
### Release note
```release-note
[platform] Fixed namespace PodSecurity label race condition when multiple Packages share a namespace
```
Instead of using per-Package SSA field owners (which is a workaround relying
on SSA mechanics), properly resolve whether a namespace should be privileged
by iterating all PackageSources and their active Packages. A namespace gets
the privileged PodSecurity label if ANY Package has a component with
privileged: true installed in it.
This fixes the race condition where Packages sharing a namespace (e.g.
linstor and linstor-scheduler in cozy-linstor) would overwrite each other's
labels depending on reconciliation order.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
When multiple Packages share a namespace (e.g. cozystack.linstor and
cozystack.linstor-scheduler both use cozy-linstor), the shared SSA field
owner "cozystack-package-controller" caused a race condition: the last
Package to reconcile would overwrite the namespace labels set by others.
This meant that if cozystack.linstor set pod-security.kubernetes.io/enforce=
privileged and then cozystack.linstor-scheduler reconciled without the
privileged flag, SSA would remove the label. On Talos clusters (which
default to "baseline" PodSecurity enforcement), this caused privileged
satellite pods to be rejected with PodSecurity violations.
Fix by using per-Package field owners (cozystack-package-{name}), so each
Package independently manages its own namespace labels without interfering
with other Packages sharing the same namespace.
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
The mysql chart actually deploys MariaDB via mariadb-operator, but was
incorrectly named "mysql". Rename all references to use the correct
"mariadb" name across the codebase.
Changes:
- Rename packages/apps/mysql -> packages/apps/mariadb
- Rename packages/system/mysql-rd -> packages/system/mariadb-rd
- Rename platform source and bundle references
- Update CRD kind from MySQL to MariaDB
- Update RBAC, e2e tests, backup controller tests
- Keep real MySQL CLI/config tool names unchanged (mysqldump, [mysqld], etc.)
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
<!-- Thank you for making a contribution! Here are some tips for you:
- Start the PR title with the [label] of Cozystack component:
- For system components: [platform], [system], [linstor], [cilium],
[kube-ovn], [dashboard], [cluster-api], etc.
- For managed apps: [apps], [tenant], [kubernetes], [postgres],
[virtual-machine] etc.
- For development and maintenance: [tests], [ci], [docs], [maintenance].
- If it's a work in progress, consider creating this PR as a draft.
- Don't hesistate to ask for opinion and review in the community chats,
even if it's still a draft.
- Add the label `backport` if it's a bugfix that needs to be backported
to a previous version.
-->
## What this PR does
### Release note
<!-- Write a release note:
- Explain what has changed internally and for users.
- Start with the same [label] as in the PR title
- Follow the guidelines at
https://github.com/kubernetes/community/blob/master/contributors/guide/release-notes.md.
-->
```release-note
[
- add BackupClass Velero and strategy for VM
- dashboard fixes for BackupClass and simplifed Plan/BackupJob API
]
```
