Commit Graph

618 Commits

Author SHA1 Message Date
Andrei Kvapil
126a842dde [kubernetes] Pre-install NVIDIA GPU Driver
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-24 00:38:53 +02:00
Timofei Larkin
f1624353ef Hash tenant config and store in configmap
Every tenant now creates a configmap in its __tenant__ namespace with a
sha256 of its values. Tenants (and eventually all other apps), watch the
configmap in their __release__ namespace, by referencing it in the
valuesFrom part of the HelmRelease. `tenant-root` is an exception, since
it is the only tenant where the release namespace is the same as the
tenant namespace. It references a different configmap in its valesFrom,
created and reconciled by the cozystack installer script. Part of #802.

Signed-off-by: Timofei Larkin <lllamnyp@gmail.com>
2025-04-22 18:57:18 +02:00
kklinch0
719fdd29cc [platform]: fix migrations
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-22 17:40:59 +03:00
klinch0
ccedcb7419 [kubernetes] Fix tenant addons removal (#835)
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Expanded the pre-delete operation to target additional components,
including cert-manager and vertical pod autoscaler resources.
- **Chores**
- Updated chart version to 0.18.1 and revised version mappings for
improved tracking.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-22 16:07:54 +03:00
Andrei Kvapil
80576cb757 [platform]: add VerticalPodAutoscaler for Cozystack dashboard (#828)
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

- **New Features**
- Introduced automated resource management for dashboard components
using Kubernetes VerticalPodAutoscaler, enabling dynamic adjustment of
CPU and memory resources.
- **Chores**
- Updated configuration to explicitly set resource presets to "none" for
dashboard, frontend, and related components.
- Added a migration script to ensure Keycloak configuration is properly
reconciled in managed environments.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-22 12:44:27 +02:00
kklinch0
fde6e9cc73 [platform]: add migration for kube-rbac-proxy daemonset
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-22 13:05:48 +03:00
Timofei Larkin
57ca60c5a5 [platform] Fix installing HelmReleases on initial setup (#833)
fixes https://github.com/cozystack/cozystack/issues/832

This PR fixes regression on installing helmreleases, also some refactor

Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-22 14:01:32 +04:00
Andrei Kvapil
1d0ee15948 [kubernetes] Fix tenant addons removal
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-22 11:42:40 +02:00
Andrei Kvapil
be84fc6e4e Fix: installing HelmReleases on initial setup
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-22 09:48:53 +02:00
kklinch0
73a3f481bc (platform): make lower resource request for capi-kamaji-controller-manager
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-18 15:00:52 +03:00
kklinch0
4e9446d934 [monitoring] fix vpa for vmagent delete resources
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-17 21:38:28 +03:00
Andrei Kvapil
86210c1fc1 Release v0.30.2 (#813)
This PR prepares the release `v0.30.2`.
(Please merge it before releasing draft)
2025-04-16 09:45:47 +02:00
kvaps
e96f15773d Prepare release v0.30.2
Signed-off-by: kvaps <kvaps@users.noreply.github.com>
2025-04-15 07:42:59 +00:00
klinch0
bc5635dd8e [monitoring] add vpa for users k8s clusters (#806)
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **Chores**
- Updated the application version to 0.18.0 with refined version
tracking for improved deployment clarity.
  
- **New Features**
- Enhanced the monitoring agents integration with updated dependency
management.
- Introduced new deployment configurations for the vertical pod
autoscaler and its custom resource definitions, offering customizable
override options and improved reconciliation strategies.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-15 09:38:38 +02:00
Andrei Kvapil
05d6ab9516 [platform] Another logic for deleting components
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-14 17:02:50 +02:00
kklinch0
5a5cf91742 (platform): revert API_VERSIONS_FLAGS
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-14 15:36:16 +03:00
klinch0
6a0d4913f2 [platform] fix deleting bundles (#809)
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Enhanced the container image with an additional YAML processing tool
for improved configuration management.
- Introduced new workflow commands that streamline deployment operations
by reconciling resource changes and automating cleanup.
- Enabled management of disabled components by automatically suspending
and flagging inactive deployments for optimized system performance.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-14 14:28:08 +03:00
klinch0
685e50bf6c [monitoring] add vpa for users k8s clusters (#806)
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **Chores**
- Updated the application version to 0.18.0 with refined version
tracking for improved deployment clarity.
  
- **New Features**
- Enhanced the monitoring agents integration with updated dependency
management.
- Introduced new deployment configurations for the vertical pod
autoscaler and its custom resource definitions, offering customizable
override options and improved reconciliation strategies.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-14 14:07:35 +03:00
kklinch0
f90fc6f681 [platform] fix deleting bundles
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-14 13:22:33 +03:00
kklinch0
da8100965f [monitoring] add vpa for users k8s clusters
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-11 14:52:26 +03:00
kvaps
7f4af5ebbc Prepare release v0.30.1
Signed-off-by: kvaps <kvaps@users.noreply.github.com>
2025-04-11 10:07:16 +00:00
Andrei Kvapil
8f575c455c [monitoring] Refactor management etcd monitoring config (#799)
* Reuse the vmagent's serviceaccount
* Mount the serviceaccount token instead of manually creating secrets
* Give the kube-rbac-proxy a unique labelset to avoid targeting wrong
pods

Resolves #789 

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Introduced a new proxy component within the monitoring configuration.

- **Refactor**
  - Updated resource labeling for improved consistency.
- Revised service account references and authentication settings for a
more streamlined operation.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-11 10:57:30 +02:00
Andrei Kvapil
819166eb35 [monitoring] create a new version to include fix for vlogs image (#803)
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

- **Chores**
  - Updated the monitoring application version from 1.9.1 to 1.9.2.
- Refined version reference details by shifting from a dynamic reference
to a fixed commit identifier for improved tracking and reliability.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-11 10:55:54 +02:00
kklinch0
f507802ec9 (monitoring) patch for fix vlogs image
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-11 11:51:33 +03:00
Timofei Larkin
60b96e0a62 Refactor management etcd monitoring config
* Reuse the vmagent's serviceaccount
* Mount the serviceaccount token instead of manually creating secrets
* Give the kube-rbac-proxy a unique labelset to avoid targeting wrong
  pods

Signed-off-by: Timofei Larkin <lllamnyp@gmail.com>
2025-04-10 16:59:43 +03:00
kvaps
1d377bab9d Prepare release v0.30.0
Signed-off-by: kvaps <kvaps@users.noreply.github.com>
2025-04-10 12:50:47 +00:00
Andrei Kvapil
799690dc07 [ci] Allow tests to run in paralel (#786)
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 14:36:01 +02:00
Andrei Kvapil
aa02d0c5e6 [tests] fix tests and add retry for monitoring hr tests (#791)
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Introduced explicit versioning for monitoring components, now using
version *v1.17.0-victorialogs*.
- Enhanced the Docker setup by including Flux CD support for improved
deployments.

- **Chores**
- Optimized deployment automation to streamline service readiness checks
and improve overall reliability.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-10 14:35:50 +02:00
Andrei Kvapil
6b8ecf3953 Upd: Kube-OVN to v1.13.8 (#797) 2025-04-10 14:35:22 +02:00
Andrei Kvapil
e5b81f367e Upd: Cilium to v1.17.2 (#796)
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 14:35:07 +02:00
Andrei Kvapil
ba4798464d Upd: Kamaji to edge-25.3.2 (#795)
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 14:34:50 +02:00
Andrei Kvapil
655e8be382 Upd: Keycloak-operator to v1.25.0 (#794)
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Upgraded to a new release version, offering enhanced integration and
secure client configuration options.
- Expanded realm settings now support advanced user profile
customization and robust email configuration for streamlined operations.
- Improved administrative views deliver clearer insights for managing
your system.

- **Documentation**
- Installation and release details have been updated to accurately
reflect the latest version.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-10 14:34:32 +02:00
Andrei Kvapil
fd9a5b0d7b Update Cluster-API operator to v0.18.1 (#793)
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Upgraded the Cluster API Operator release to version 0.18.1 with an
updated manager image.
- Introduced a new configuration option that enables conditional
deployment hooks, allowing for custom post-install and post-upgrade
actions.
- Enhanced resource synchronization with dynamic feature gate settings,
providing more flexible and coordinated deployments.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-10 14:34:16 +02:00
Andrei Kvapil
371791215a Upd: Kube-OVN to v1.13.8
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 14:32:34 +02:00
Andrei Kvapil
4c220bb443 Upd: Cilium to v1.17.2
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 14:28:51 +02:00
Andrei Kvapil
e27611d45a Upd: Kamaji to edge-25.3.2
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 14:28:09 +02:00
Andrei Kvapil
a3e647c547 Upd: Keycloak-operator to v1.25.0
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 14:26:58 +02:00
Andrei Kvapil
be52fe5461 Update Cluster-API operator to v0.18.1
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 14:26:09 +02:00
kklinch0
1d639fda0d [tests] fix tests and add retry for monitoring hr tests
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-10 15:23:31 +03:00
Andrei Kvapil
bbdde79428 Upd: victoria-metrics operator to v0.55.0
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 14:23:12 +02:00
Andrei Kvapil
1966f86120 [ci] Allow to run tests in paralel
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 11:35:06 +02:00
Andrei Kvapil
cc9abfe03f [ci] Add talos-kernel and talos-initramfs to assets
Signed-off-by: Andrei Kvapil <kvapss@gmail.com>
2025-04-10 10:54:54 +02:00
Andrei Kvapil
9e6478b9c9 [linstor] Add plunger check for disconnected DRBD peers. (#707)
Sometimes DRBD devices get stuck in "Connecting" state, probably due to
some
race conditions. This scriptlet provides a workaround for such
situations.
2025-04-10 10:38:29 +02:00
Andrei Kvapil
3a295c4474 Add guard against empty cloudInit in vm-instance app (#646)
Prevent the VM resource from referencing a non-existent secret when
`sshKeys` are set and `cloudInit` is set to empty.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Improved cloud-init configuration handling with conditional logic and
clearer error messaging when expected configuration values are missing.

- **Documentation**
- Refined virtual machine configuration guides by reformatting parameter
tables and correcting typographical errors in parameter descriptions.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-10 10:37:50 +02:00
klinch0
2966922c0b feat(vpa): separate-crds (#781)
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

- **New Features**
- Improved autoscaling deployment by integrating an additional component
for managing custom resource definitions.
- Enhanced dependency management now ensures critical prerequisites are
deployed in the correct order.
- Introduced an automated update mechanism to keep resource definitions
current.
- Added a new configuration option, giving users the flexibility to
enable or disable custom resource definitions as needed.
- Introduced two new Custom Resource Definitions:
`VerticalPodAutoscalerCheckpoint` and `VerticalPodAutoscaler`.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2025-04-10 11:35:36 +03:00
Denis Seleznev
991c7e1943 Handle empty cloudInit.
Add a no-op user-data when sshKeys are specified.

Signed-off-by: Denis Seleznev <kto.3decb@gmail.com>
2025-04-10 10:07:43 +02:00
kklinch0
c31a7710ad feat(vpa): separate-crds
Signed-off-by: kklinch0 <kklinch0@gmail.com>
2025-04-10 10:57:50 +03:00
Andrei Kvapil
f4cace093c Add a setting to VMs that allows users to trigger cloud-init full reconfiguration. (#767)
This will trigger cloud-init reinitialization, including ssh keys update
and static network config refresh.
2025-04-10 09:20:55 +02:00
Denis Seleznev
01e417d436 Add Linstor plunger scriptlet to fix DRBD devices that are stuck disconnected.
Sometimes DRBD devices get stuck in "Connecting" state, probably due to some
race conditions. This scriptlet provides a workaround for such situations.

Signed-off-by: Denis Seleznev <kto.3decb@gmail.com>
2025-04-10 03:49:23 +02:00
Denis Seleznev
261ce4278f Add a setting to VMs that allows users to trigger cloud-init full reconfiguration.
Changing `cloudInitSeed`  will trigger cloud-init reinitialization, including ssh keys update and static network config refresh.

Signed-off-by: Denis Seleznev <kto.3decb@gmail.com>
2025-04-09 20:48:18 +02:00