rancher/fleet v0.12.0-beta.2 on GitHub

Notes

This release adds a new HelmOps controller that allows you to manage Helm charts using the HelmApp CRD. This controller is experimental and is not enabled by default.
Several log messages have been improved to provide more context and information about the errors that occurred. The conditions in the status fields of resources are used by the UI to display warnings and we are continuously improving them.

When a bundle fails to deploy to a cluster, e.g. because the cluster is missing a required label that is to be used as a value, the bundle will now show the affected cluster in the error message. This makes it easier to identify which clusters are affected by the error.

The agent has been switched from a StatefulSet to a Deployment. This change allows the agent to scale horizontally and is a step towards making the agent more resilient to failures. The agent now exposes the number of worker goroutines it is running, which can be used to tune the agent's performance.

Additions

Fall back to Rancher-configured CA bundles by @weyfonk in #3233
Add experimental HelmOps controller. by @0xavi0 in #3092
Add benchmark suite by @manno in #3120
Add templateValues to HelmApp, Bundle, etc by @Danil-Grigorev in #3267
Add jitter and resync to polling by @manno in #3151
Adds tolerations from fleet-controller's deployment to local agent and git job by @0xavi0 in #3362
Propagate tolerations from values to cleanup jobs by @weyfonk in #3281
Switch agent from StatefulSet to Deployment by @p-se in #3364
Expose agent worker counts in fleet chart by @weyfonk in #3152
Make max concurrent reconciles configurable for agent and remaining controllers by @weyfonk in #3094
Use a delaying enqueue handler for drift detection by @aruiz14 in #3401
Support ignoring resources in bundle diffs by @weyfonk in #3425
Fleet apply retries bundle creation on conflicts by @0xavi0 in #3407

Logging and Status Conditions:

Add extra error info when downloading remote charts by @0xavi0 in #3368
Adds message context to fleet apply errors by @0xavi0 in #3385
Prefix log fields for resource conflicts by @manno in #3242
Add affected cluster to template error message by @p-se in #3278
Add template errors to bundle and GitRepo status by @p-se in #3114
Adds logs for new commit or error checking for the latest by @0xavi0 in #3182
Update gitcloner/fetcher error messages for SCP/SSH GitRepo URLs by @manno in #3430

Resources in Status Fields:

Calculate Clusters resourceCount from BundleDeployments instead of GitRepos by @aruiz14 in #3102
Add resourceCounts and incomplete state to BundleDeploymentStatus by @aruiz14 in #3203
Add PerClusterResourceCounts to GitRepo status by @aruiz14 in #3209
Rework how resources are calculated by @aruiz14 in #3238
Simplify PerClusterState and include all states and cluster IDs by @aruiz14 in #3287

Bugfixes

Deduplicate status messages by @p-se in #3042
Remove reconcile error from retry by @manno in #3045
Fix namespace target customization support with no defaults by @weyfonk in #3052
Apply defaults from gitrepo restrictions by @manno in #3056
Disable jobs cleanup cronjob if GitOps is disabled by @weyfonk in #3130
Fix fleet-agent chart validation by @weyfonk in #3150
Prevent empty Patch updates by @aruiz14 in #3245
Change ImageScanCommit in GitRepo spec to a pointer by @0xavi0 in #3302
Always returns result with RequeueAfter set by @0xavi0 in #3307
Update Git job permissions when GitRepo changes by @p-se in #3305
Retry reconcile when we cannot add finalizer to content resource by @manno in #3310
Fix default state for GitRepo resources by @aruiz14 in #3329
Using a webhook sets the intervall to one hour again by @manno in #3337
Bundlereader needs to update OCI storage bundles, too by @manno in #3389
Fix BundleDeployment resource counts when a resource is missing by @aruiz14 in #3408
Delete both, statefulset and deployment on agent update by @p-se in #3449

What's Changed

Remove logrus message for finalizer handling by @manno in #3243
Upload Docker images to Prime registry and sign them by @thardeck in #3261
Fix typo on regristry by @sestegra in #3321
Bump k8s modules to 1.32 and Helm to 3.17.0 by @manno in #3317
Add debug logging around drift correction and disable test by @manno in #3334
Incorporate kube-version and rancher-version annotations in chart by @thardeck in #3352
Use auto bump functionality in release charts by @thardeck in #3365
Move agent registration into controller container by @p-se in #3416
Move agent's clusterstatus controller into the controller container by @p-se in #3429
Remove "fleet apply" config files from bundle resources by @manno in #3403

New Contributors

@sestegra made their first contribution in #3321

Full Changelog: v0.11.5...v0.12.0-beta.2