The `cert-issuer-name` and `cert-issuer-kind` properties point an app's generated `Certificate` at a cert-manager issuer created outside of Dokku, allowing certificates to be issued through solvers the built-in letsencrypt integration cannot use, such as `dns01` for wildcard certificates. Setting an issuer enables https on its own, as a manually managed issuer has no email for Dokku to configure. An imported certificate still takes precedence, and `letsencrypt-server false` remains the single off switch. Dokku warns before a build starts when the referenced issuer is absent from the cluster, without blocking the deploy. Wildcard domains no longer collide with their apex domain when generating ingress names, and `letsencrypt-server` values are now validated when set rather than at deploy time.
Kubernetes 1.20 deprecated the node-role.kubernetes.io/master label and replaced it with node-role.kubernetes.io/control-plane, and 1.24 dropped the legacy label entirely. The scheduler-k3s:cluster:add join path still filtered control-plane nodes with master=true, so on any k3s release built from Kubernetes 1.24+ the lookup matched zero nodes and aborted with "no nodes found in the cluster" before the worker could join.
Three plugins still leaked the built-in default through `<plugin>-global-<property>` after the broader raw/computed split in #8640. External tooling reading `:report --format json` for drift detection could not distinguish "set to default" from "never set" - after a `:set --global` followed by an unset, the JSON still reported the default value. The `<plugin>-global-<property>` keys now hold the raw stored value (empty when nothing has been set) and `<plugin>-computed-<property>` keys hold the effective value with per-app, global, and built-in default fallback. Every global property now has a computed sibling for shape consistency, including those whose default is empty. Internal callers in `scheduler-k3s` that previously read the default-applied global helpers now consume new `getComputed*` helpers, keeping deploy-time behavior unchanged.
Bundling these Secrets in the app helm chart caused two bugs in the scheduler-k3s plugin: a chart rollback could delete Secrets that older ReplicaSets still referenced by exact timestamped name (`env-{app}.{ts}` and `ims-{app}.{ts}`), hard-crashing pods until manual intervention; and the strategic-merge `patchMergeKey` on `imagePullSecrets` let stale entries leak into the live Deployment until the list pointed at many nonexistent Secrets. Each Secret now lives in its own helm release with a stable name (`config-{app}` and `pull-secret-{app}`), installed before the app chart on every deploy. The deployment trigger also prunes any leaked `imagePullSecrets` entries from the live Deployment so the next deploy lands on a clean list, and the rename and destroy paths uninstall the new releases (and the previously-leaked TLS release on rename) under the old app name.
Some helm chart upgrades require side-effects that a plain `helm upgrade` cannot perform, such as deleting deployments whose immutable selectors changed. Charts may now register pre and post upgrade hooks against a target version; applicable hooks fire in ascending semver order, each bracketed around an upgrade to its version, with a final upgrade to the chart's configured version when needed. The new `keda-add-ons-http` 0.12.2 hook deletes the chart-managed deployments before the upgrade so the new selectors take effect cleanly.
Previously, it was possible to have a kubernetes version mismatch on new worker nodes if a newer version of kubernetes was released prior to upgrading the master nodes to that version. This can sometimes cause cluster instability.
This change fetches the lowest node version associated with a master node and then uses that to install k3s on the worker node.
Previously, a container that exited immediately would just show an error message, which is probably not what we want.
If the pod has exited, we will display the logs on stdout for the user and then exit appropriately.
This change also refactors how log streaming works so that we respect exiting when the stream has no content (such as when the container has exited). Previously, the container would just hang around forever due to not being properly canceled. Due to this refactor, the prefix is now split out into a separate writer.
If the scheduler is not docker-local, we will assume that a separate plugin trigger will pull access/error logs on our behalf. For k3s, the naive implementation is to pull logs from a single container (multiple ingress-nginx pods will be supported in a future release).
Closes#7268
Domains cannot be reused across ingress objects, and port maps are a function of a domain, so the inversion here makes most sense (at the cost of some downtime when migrating.
This will provide the possibility for users to talk to existing kubernetes clusters, thereby removing one of the biggest reasons for having the old scheduler-kubernetes plugin around.
Different options get applied to each depending on what is needed for the role.
This also updates how we initialize the cluster to better support how metrics exposure and cross-server networking.
The cron id is too long to store in the cronjob, so we generate a short identifier and look that up each time on deploy to ensure stability across time.
