Container Monitoring 12.0.0

Resources in the lm-container now display a monitoringMode labelc argus.monitoring-mode that indicates the active monitoring mode. The label value reflects the mode configured in the lm-container.

The lm-container Helm chart no longer uses the Helm lookup function to dynamically fetch Kubernetes resources (for example, Services or Secrets) at template render time. This enables seamless installation on tools which use helm template for installation.

The lm-container Helm charts now use kube-state-metrics version 6.3.0, providing enhanced metric coverage and compatibility with newer Kubernetes versions.

Argus now supports deployment with a custom ServiceAccount. This enhancement enables improved flexibility for integrating with existing cluster policies and CI or CD setups. You can specify your own ServiceAccount through deployment configurations to align Argus with organizational security practices.

Argus logs now include throttling messages that indicate when resources are being throttled by the platform. This enhancement increases visibility into platform-level throttling behavior, helping you identify when the platform has reached its limit and can no longer accept new devices.

Added security updates to the internal components of LM Container.

You can now configure the priorityClassName property for the lm-logs DaemonSet directly through Helm values. This ensures the lm-logs pods receive the appropriate scheduling priority and are not preempted on EKS nodes.

In addition, lm-logs now support SystemD (journald) configuration, enabling FluentD to collect node-level logs on Amazon Linux 2023 and other journald-based operating systems.

The Argus and Collectorset-Controller batch pods can enter an Error state during execution. This issue is resolved in release 12.0.1.