v0.46.0

Features

• EventSource can now match CloudEvents received over NATS/JetStream. More here and here
• Lua can be used to generate at run time what needs to be deployed. It is an alternative to go templating.  More here
• Sveltos, by default, deploys Helm charts and resources in the order they appear in a ClusterProfile/Profile, retrying on errors
  without proceeding to subsequent deployments. For example, if a ClusterProfile lists three Helm charts and the second fails to deploy,
  Sveltos will not attempt to deploy the third. However, enabling the Spec.ContinueOnError setting allows Sveltos
  to proceed with deploying the third Helm chart and then retry the failed second chart. More here
• When registering the management cluster, Instead of generating a token with an expiration, the token is taken from a Secret of type kubernetes.io/service-account-token. More here

Bug Fixes

• When running in agent less mode (Sveltos agent and drift detection deployed in the management cluster for each managed cluster), those agent t receive the managed cluster's kubeconfig from a Secret. These kubeconfigs can expire (e.g., GKE tokens have a maximum lifespan of 48 hours). Sveltos includes a mechanism to proactively renew these tokens. The SveltosCluster controller can be configured to periodically refresh tokens before they expire, preventing disruptions. However the drift-detection controller and Sveltos-agent when deployed in the management cluster, lacked the ability to retrieve an updated kubeconfig. Consequently, upon kubeconfig
  expiration, the controller encountered numerous authorization errors, effectively ceasing operation. More here and here
• Fix an issue fetching helm charts from not OCI-compliant private registries. More here