Skip to content

🔀 Merge develop into master (Release)#559

Merged
vshnbot merged 19 commits intomasterfrom
develop
Dec 16, 2025
Merged

🔀 Merge develop into master (Release)#559
vshnbot merged 19 commits intomasterfrom
develop

Conversation

@github-actions
Copy link
Contributor

Auto-generated PR to release from develop to master.

mdnix and others added 19 commits December 2, 2025 14:32
🔀 Merge develop into master (Release)
K8s doesn't allow any `_` in object names. However `_` is allowed for
user names in database systems. These differences have to be escaped
when applying the objects to the API.

However the newly introduced `IsResourceRead()` function did not account
for that. Which resulted in resources not being found. This in turn
caused Crossplane to remove the supposedly non-existing resources from
the API.

Additionally this commit will set all the objects to be applied with the
policy to only create and update the data. Never to delete it. Deletion
will have to be a manual and deliberate operation from now on. This also
ensures feature parity with the PostgreSQL implementation.
Make sure names are escaped during ready check
This sets the necessary kubebuilder flags in the service CRDs to allow
setting the number of instances to 0.
If the vshnmariadb instance is scaled to 0, we also return early and
dont create proxysql.
This allows suspending an instance even if the service level is set to
guaranteed availability. Since we might need to suspend an instance for
a customer, this has to be possible.
The replicaCount from the claim was never passed to the actual helm
values. This fixes that.
This disables the cronjob for vshnnextcloud if an instance has 0
replicas.
This passes the instances field to the replicaCount in the helm values
for forgejo.
This ensures that collabora is also suspended alongside the main
nextcloud pod if instances = 0
Ensure that vshnforgejo can't be scaled beyond 1 replica for now since
we don't support HA yet.
CNPG can't handle 0 replicas for a cluster. It instead has a hibernation
feature. The hibernation procedure will delete the primary Pod and then
the replica Pods, avoiding switchover, to ensure the replicas are kept
in sync.
This checks if an instance is suspended an stops the SLI exporter from
probing those instances prevent false alerts.
This removes the backup schedules and suspends database operations like
repack and vacuum for suspended instances.
Allow instances to be downscaled to 0
The nextcloud helmchart 8.6.0 uses seperate registry and repository
configurations, which doesn't work with the previous implementation.
Update Nextcloud Helmchart to 8.6.0
This adds the necessary storage class annotations for APPUiO on
Exoscale.
Add Exoscale specific annotations for quotas
@vshnbot vshnbot merged commit a4578cd into master Dec 16, 2025
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants