KubeSphere Enterprise v3.5.0 Release Notes

DevOps

Enhancements & Updates

• Support for customizing the Jenkins service types in the Installer configuration, as well as optimizing the timeouts of readiness and liveness probe for the ArgoCD service.
• Merge DevOps-related alerting rules into a group.
• Upgrade Jenkins to version 2.346.3, and upgrade some plugins.
• Add a plugin to get the build user information in Jenkins and add this step to the graphical pipeline panel.
• Refactor the DevOps interface to avoid getting data directly from jenkins.

Bug Fixes

• Fix the request branch interface bug of DevOps single branch pipeline.
• Fix the discover branches default value for git type of DevOps multi-branch pipeline.
• Fix the issue that shell commands cannot be saved when editing the pipeline graphically.
• Fix the issue that the pipeline run log is too large to view online.
• Fix the error when replaying a pipeline.
• Fix the layout of the DevOps runlog page.
• Fix the display issue of DevOps projects under Recent Acess.

Multi-cluster

Enhancements & Updates

• Validate cluster roles when adding clusters.

Bug Fixes

• Fix webhook validation issue for new clusters.
• Clean up clusters only when a cluster is ready.
• Fix the issue that cluster authorization is not revoked after deleting a cluster.
• Fix the issue that the added tag is not shown after adding a cluster.

Observability

Features

Monitoring

• Add the calico-exporter component to provide monitoring metrics for Calico, including BGP, IPPool, and IPAM.

Alerting

• Add support for filtering alerts by keywords.
• Support adding labels when adding/updating rules.

Logging

• Export logs in positive or reverse order.

Notification

• Update notification-manager to support filtering notifications using regular expressions.
• Support multiple receivers, and regular expressions for notification filtering/silencing.

Enhancements & Updates

Monitoring

• Add Whizard server params in the ClusterConfiguration file.
• Optimize process-exporter configuration to improve its responsiveness.
• Update blackbox-exporter to version 0.24.0.
• Change the severity of alerting rules to info if it is none or empty.
• Update Whizard to version 0.9.0.
• At least one cluster should be selected when selecting clusters in Whizard.
• Remove the "Usage" information from "Cluster Quota Statistics" on the overview page of a cluster.

Alerting

• Increase CPU limit for default-http-backend to avoid CPUThrottlingHigh alerts.
• Add a NodeExporterDown rule to the built-in node-exporter GlobalRuleGroup, and configure kube-state-metrics to export kube_node_info metric by default.
• Increase CPU limit for some monitoring components to avoid CPUThrottlingHigh alerts.
• Support configuring alerting rules at the Info level.
• Add calico-related alerting rules.
• Show full names of monitoring targets in alerts.

Logging

• Optimize parameters of fluentbit to improve log collection efficiency.
• Support automatic update of usernames and passwords of external Elasticsearch or OpenSearch clusters.
• Add a switch to turn on/off Elasticsearch/OpenSearch curator.
• Add a switch to disable output logs to Elasticsearch/OpenSearch.

Notification

• Update notification-manager to v2.4.0.

Auditing

• Update kube-auditing to v0.3.0.

Bug Fixes

Monitoring

• Fix the storage statistics issue of workspaces.
• Fix the scope of rules when configured as template rules in GlobalRuleGroups.
• Fix the issue that some target down rules cannot alert as expected in multi-cluster.
• Optimize the PVC monitoring prompt.

Alerting

• Fix the issue that when a member cluster is removed, alert messages from that member cluster are still shown in Whizard.
• Fix alerting rules created for a cluster still exist when the cluster is removed.
• Update alerts statistics promQL.
• Fix unexpected duplicate notifications when the notification channel is configured with a cluster filter in multi-cluster.
• Fix the issue that cannot get the system-workspace of a cluster when creating a custom alerting policy.
• Fix the issue that labels are required when creating an alerting policy.
• Fix the issue that rule names are not shown in alerting messages under clusters and projects.
• Fix the issue that project names are not filled automatically in the monitoring target of alerting rules.

Notification

• Fix the missing notification history issue when the container runtime is contained.

Storage

Features

• Storage usage can be categorized by storage type.

Enhancements & Updates

• Remove GlusterFS and Ceph RBD options when creating storage classes.

Bug Fixes

• Fix the issue that storage class authorization rules are not working in some cases.
• Fix storageclass-accessor webhook rules.

Gateway

Bug Fixes

• Fix the checksum error when updating the workspace gateway if no gateway namespace is configured.
• Fix the permission error when viewing the workspace gateway by an account with regular or self-provisioner permission on a workspace.
• Fix the permission error when viewing the monitoring info of the workspace gateway by an account with regular or self-provisioner permissions on a workspace.

DMP

Enhancements & Updates

• OpenSearch: add Logstash configuration and standalone monitoring and alerts.
• OpenSearch: upgrade from 2.3.0 to 2.8.0.
• Kafka: support user-specific flow and speed limits, and add description in permission rules.
• Kafka: the default certificate validity is adjusted to 10 years.
• Kafka: support increasing the number of partitions in a created topic.
• Kafka: add the Jmx parameter, which supports opening the Jmx port for cmak.
• Upgrade Kafka Operator to version 0.2.0 and OpenSearch Operator to version 2.0.25, remove ECK and ClickHouse apps, and remove Redis Cluster & Standalone 6.2.5.
• OpenSearch, Kafka, and Redis support preset cluster alerts.

Bug Fixes

• OpenSearch: fix PromQL for some monitoring metrics.
• OpenSearch: change Logstash to Statefulset to ensure that monitoring information can still be obtained even if pods are recreated.

User Experiences

Features

• Support filtering pods by pod IP.
• Support displaying the LDAP login page via URL.
• Pods in the pod list can be filtered by pod IP.

Enhancements & Updates

• Change the default resync period of Informer to 10 hours.
• When a license expires, the notification history page and the user notification settings page are still available as before.

Bug Fixes

• Fix global alerting styles of Whizard.
• Fix the issue that the added container image is incorrect if not setting the image tag when creating a workload.
• Fix the issue that container images are unsearchable when editing workloads.
• Show status when editing crd info.
• Fix the failure when deploying from an app template without values.yaml.
• Fix the issue that the admin info is not synchronously cleared after clearing the workspace info in the window for allocating projects to workspaces.

Authentication & Authorization

Enhancements & Updates

• Add resource configuration for Redis, OpenLDAP, and metrics-server.

Network

Enhancements & Updates

• Update namespaceNetworkPolicy crd to support user-defined port ranges. (Only back-end support)