Enhancements

This section outlines new features and enhancements introduced in the Cluster release 6.18.0.

Graceful MCR upgrade

Implemented a graceful Mirantis Container Runtime (MCR) upgrade from 19.03.14 to 20.10.5 on existing Container Cloud clusters.

Learn more

MKE logs gathering enhancements

Improved the MKE logs gathering by replacing the default DEBUG logs level with INFO. This change reduces the unnecessary load on the MKE cluster caused by an excessive amount of logs generated with the DEBUG level enabled.

Log verbosity for StackLight components

Implemented the capability to configure the verbosity level of logs produced by all StackLight components or by each component separately.

Learn more

StackLight configuration parameters: Log verbosity

StackLight log level severity setting in web UI

Implemented the capability to set the default log level severity for all StackLight components as well as set a custom log level severity for specific StackLight components in the Container Cloud web UI. You can update this setting either during a managed cluster creation or during a post-deployment configuration.

Learn more

Improvements to StackLight alerting

Implemented the following improvements to StackLight alerting:

  • Added the following alerts:

    • PrometheusMsTeamsDown that raises if prometheus-msteams is down.

    • ServiceNowWebhookReceiverDown that raises if alertmanager-webhook-servicenow is down.

    • SfNotifierDown that raises if the sf-notifier is down.

    • KubeAPICertExpirationMajor, KubeAPICertExpirationWarning, MKEAPICertExpirationMajor, MKEAPICertExpirationWarning that inform on SSL certificates expiration.

    • KubeContainersCPUThrottlingHigh that raises in case of containers CPU throttling.

    • KubeletDown that raises if kubelet is down.

  • Removed the following inefficient alerts:

    • PostgresqlPrimaryDown

    • FileDescriptorUsageCritical

    • KubeCPUOvercommitNamespaces

    • KubeMemOvercommitNamespaces

    • KubeQuotaExceeded

    • ContainerScrapeError

  • Reworked a number of alerts to improve alerting efficiency and reduce alert flooding.

  • Reworked the alert inhibition rules to match the receivers.

  • Updated Alertmanager to v0.22.2.

  • Changed the default behavior of the Salesforce alerts integration. Now, by default, only Critical alerts will be sent to the Salesforce.

Learn more

StackLight node labeling improvements

Implemented the following improvements for the StackLight node labeling during a cluster creation or post-deployment configuration:

  • Added a verification that a cluster contains minimum 3 worker nodes with the StackLight label for clusters with StackLight deployed in HA mode. This verification applies to cluster deployment and update processes. For details on how to add the StackLight label before upgrade to the latest Cluster releases of Container Cloud 2.11.0, refer to Upgrade managed clusters with StackLight deployed in HA mode.

  • Added a notification about the minimum number of worker nodes with the StackLight label for HA StackLight deployments to the cluster live status description in the Container Cloud web UI.

Caution

Removal of the StackLight label from worker nodes along with removal of worker nodes with StackLight label can cause the StackLight components to become inaccessible. It is important to keep the worker nodes where the StackLight local volumes were provisioned.

Learn more

Upgrade managed clusters with StackLight deployed in HA mode

Salesforce feed update

Implemented the capability to enable feed update in Salesforce using the feed_enabled parameter. By default, this parameter is set to false to save API calls.

Learn more

Proxy configuration on existing clusters

Implemented the capability to add or configure proxy on existing Container Cloud managed clusters using the Container Cloud web UI.

Learn more

Change a cluster configuration

Ceph Octopus

Upgraded Ceph from 14.2.19 (Nautilus) to 15.2.13 (Octopus) and Rook from 1.5.9 to 1.6.8.

Documentation enhancements

On top of continuous improvements delivered to the existing Container Cloud guides, added the following procedures:

Hyperconverged Ceph improvement

Technology Preview

Implemented the capability to define Ceph tolerations and resources management through the KaaSCephCluster spec using the native Rook way for setting resource requirements for Ceph daemons.

Learn more

Operations Guide: Enable Ceph tolerations and resources management

Ceph cluster status

Improved the MiraCephLog custom resource by adding more information about all Ceph cluster entities and their statuses. The MiraCeph, MiraCephLog statuses and MiraCephLog values are now integrated to KaaSCephCluster.status and can be viewed using the miraCephInfo, shortClusterInfo, and fullClusterInfo fields.

Learn more

Ceph Manager modules

Implemented the capability to define a list of Ceph Manager modules to enable on the Ceph cluster using the mgr.modules parameter in KaaSCephCluster.

Learn more

MOSK documentation: Ceph advanced configuration - General parameters