Services

Explore Kyndryl Bridge Services

Actionable Insights
Published On Jun 03, 2026 - 7:53 PM

Actionable Insights

Explore Actionable Insights that assist in gaining a deep understanding of cluster resource consumption and version maintenance, enabling an optimized cloud service inventory.

Introduction

Container Cluster Management provides Actionable Insights so that as a Site Reliability Engineer (SRE), you can manage your container resources effectively. These insights allow you to identify which clusters might require updates due to unsupported Kubernetes, OpenShift versions, or those nearing end-of-support. You can also pinpoint containers that are either under or over-utilizing resources by comparing actual consumption against allocated Memory and CPU usage.
Calculations for Optimal Resource Utilization
As you direct your DevOps team to establish a baseline for Memory and CPU usage, you can rely on the algorithm, which uses CPU request and limit, Memory request and limit, and historical data, to inform you of required resource adjustments as described in the following limited examples:
  1. Optimal Memory Limit Recommendations for Containers
    :
    • Metrics include memory requests, limits, frequency of Out of Memory Kills, and usage history.
    • Action: SRE reviews and adjusts values as needed.
  2. Optimal CPU Limit Recommendations for Containers
    :
    • Metrics include CPU requests, limits, and consumption history.
    • Action: SRE adjusts node sizing and plans VM migrations for load balancing.
Utilization Rules
  • Containers are underutilized if CPU or Memory usage is less than 50% of the requested amount.
  • Containers are overutilized if experiencing frequent Out of Memory issues or if Memory usage exceeds 90% of the set limit.
  • Upscaling is considered if CPU usage exceeds 90% of the set limit for significant periods.
  • Containers lacking CPU Limit or Request settings are flagged accordingly.
  • While generating Capacity & Utilization insights, the ML models consider lesser buffers for lower environments and more buffers for higher environments as follows:
    • Dev - 20%
    • UAT - 30%
    • Pre-Prod – 40%
    • Prod – 50%
A minimum of 7 days of usage data is required for generating accurate insights. Container replicas are grouped under their deployment names for resource usage analysis.

Actionable Insights page

Select any tile within the Actionable Insights widget on the main dashboard to navigate to the Actionable Insights page. The data displayed on this page depends on which tile you select from the dashboard. If, for example, you select Security & Compliance, the page will display data associated with resources that have exceeded warning or comprise thresholds for security and compliance concerns.
The page contains the following elements:
  • Bread crumbs for efficient navigation.
  • Header to identify your current location in the Bridge Kyndryl portal.
  • A dropdown list enabling you to select either the default data view or a filtered view that you saved:
    1. Filter by
      Providers
      ,
      Connections
      ,
      Applications
       and
      Environments
      ,
      Insights type
      ,
      Insights category
      ,
      Kubernetes
      , and
      OpenShift
      , and then click
      Apply
       to view only data relevant to your current requirement.
    2. Click
      Save View
       to save the filtered view for a later time, providing a relevant name when prompted.
    3. If you further filter data, you can update the current view by clicking
      Update View
      .
    4. Click
      Delete
       to delete the saved view.
  • Insights Summary, which displays four of tiles that navigate to insight details within four classes:
    • Security & Compliance
    • Performance
    • Reliability
    • Optimization
  • Insight Type
     filter that enables filtering by specific insights within the current insight class.
  • Severity
    filter that enables filtering by severity level.
  • An insights table that provides data generated for each cluster, for the current insights class (Security & Compliance, Performance, Reliability, and Optimization..

Supported Insights

Insights for Containers and Clusters running Kubernetes or OpenShift enable preemptive action. Insights are listed within the context of four insight classes:
  • Security and Compliance
  • Performance
  • Reliability
  • Optimization
Supported Insights include the following:
  • Security & Compliance
    • K8S_VERSION_REACHED_END_OF_LIFE: Need immediate action.
    • K8S_VERSION_CROSSED_END_OF_SUPPORT: K8s Version crossed support life, now only on maintenance support.
    • K8s_VERSION_EXTENDED_SUPPORT: K8s Version on extended support.
    • K8S_VERSION_NEARING_END_OF_SUPPORT: Low priority, can wait till it is critical.
    • NAMESPACE_WITHOUT_NETWORK_POLICY: Namespaces that are not secured with network policies.
  • Performance
    • CPU_RUNNING_OUT: CPU usage with respect to limits is above 90% and is forecasted to go above limits in the next 7 days.
    • CPU_OVER_UTILIZED:CPU: usage with respect to limits is above 90% but not running out.
    • MEMORY_OVER_UTILIZED: Memory usage with respect to limits is above 90% but not running out.
  • Reliability
    • POD_FAIL_CRASHLOOP: Pod is in Crashloop back off state and crashing.
    • MEMORY_RUNNING_OUT: Memory usage with respect to limits is above 90% and is forecasted to go above limits in the next 7 days.
    • POD_FAIL_PENDING: Pod is in pending state and not up.
    • POD_FAIL_RESTART: Pod has restarts, but not in crash loop back off or pending.
    • SINGLETON DESIGN: Pods with one replica.
    • PVC_OVER_UTILIZED: Persistent Volume Claim is more than 20% of what is provisioned.
    • PVC_RUNNING_OUT: Based on the current Persistent Volume Claim usage trend, provisioned PVC will run out in seven days.
  • Optimization
    • High Risk MINIMAL_CPU_MEMORY_UTILIZED: CPU and Memory usage with respect to requests is below 10%.
    • CPU_AND_MEMORY_LIMITS_NOT_SET_MINIMAL_USAGE: CPU and Memory limits are not set and CPU and Memory usage with respect to requests is below 10%.
    • CPU_REQUESTS_LIMITS_NOT_SET: CPU requests and limits are not set.
    • MEMORY_REQUESTS_LIMITS_NOT_SET: Memory requests and limits are not set.
    • CPU_LIMITS_NOT_SET: CPU limits is not set and the usage with respect to request is above 50%.
    • CPU_LIMIT_NOT_SET_UNDER_UTILIZED:
    • CPU_LIMITS_NOT_SET_MINIMAL_USAGE: CPU limits are not set and CPU and Memory usage with respect to requests is below 10%.
    • MEMORY_LIMITS_NOT_SET_MINIMAL_USAGE: Memory limits are not set and CPU and Memory usage with respect to requests is below 10%.
    • MEMORY_LIMIT_NOT_SET_UNDER_UTILIZED:
    • MEMORY_LIMITS_NOT_SET: Memory limits is not set and the usage with respect to request is above 50%.
    • CPU_UNDER_UTILIZED: CPU usage with respect to requests is below 50%.
    • MEMORY_UNDERUTILIZED
    • INACTIVE_VOLUME: The number of clusters with volumes not currently in use (not mounted or actively transferring data) but still properly recognized and managed by the system.
    • ORPHANED_VOLUME: The number of clusters with volumes that have lost association with its parent controller or application for example, the Pod or PVC was deleted, but the actual storage remains.
    • PVC_UNDERUTILIZED: Persistent Volume Claim is less than 20% of what is provisioned.
    • HPA_RECOMMENDED: Recommended minimum and maximum replicas with optimal configuration recommendation (details page).
    • NETWORK_USAGE_GROWTH:
      • High: Growth rate is above 25%.
      • Medium: Growth rate is 10% - 25%.

Actionable Insight details pages

The Resource Name in the Actionable Insights table, is a link to a page containing details for the insights data for that cluster in an expanded form.The details page presents various meta data such as Insight type, severity, application, recommended configuration, current configuration, and current usage chart.
All data is pulled dynamically and presented in near real time. The type of data varies with the current insight class.
Do you have two minutes for a quick survey?
Take Survey