Fairwinds Insights Release Notes: Kyverno Integration & GPU Metrics

Over the last several months, we’ve expanded Fairwinds Insights to give platform and operations teams deeper visibility into both policy posture and infrastructure metrics and costs. Our releases focused on enhancing the Kyverno integration and introducing GPU‑aware metrics and cost data to support better scheduling and rightsizing decisions.
Expanded Kyverno Integration in Insights
Kyverno users can now work with more complete and actionable policy information directly inside Fairwinds Insights.
Key improvements include:

Support for working with multiple Kyverno policy types (such as ClusterPolicy, Policy, and newer Kyverno policy types) so you can understand policy behavior across your clusters in a unified way.
Ingestion and display of Kyverno policies that are not managed by Insights, so unmanaged policies can appear alongside Insights‑managed policies in the same interface.
UI updates that show Kyverno policy application status and clearly label which policies are managed by Fairwinds, so users can quickly distinguish what’s driven by Insights versus managed elsewhere.
Improved handling of Kyverno admissions data, including fixes to search behavior, missing dates, and duplicated admissions, so admission events appear more consistently and can be connected to the appropriate policies and Action Items.

Taken together, these changes make it easier to operationalize Kyverno at scale: you can store Kyverno policies in Insights, map them to App Groups, view admissions and violations in one place, and easily distinguish between policies managed by Fairwinds and those managed by your own teams.
GPU Metrics
As more teams run GPU‑backed workloads for artificial intelligence (AI) and machine learning (ML), Insights now provides better visibility into how GPUs are used.
New capabilities include:

GPU‑related metrics columns in metrics tables, allowing you to see GPU information alongside CPU and memory when analyzing cluster utilization.
Ingestion of GPU metrics into the Insights back end, along with fixes to GPU‑related data migrations and support for updated aggregators, resulting in more reliable GPU time‑series data.
Cost and metrics improvements that incorporate updated aggregators and higher numeric precision, helping you analyze workloads that rely heavily on GPUs with greater confidence.

These changes make GPU usage a first‑class signal in Insights and help you understand which namespaces, clusters, or workloads are responsible for significant GPU consumption and associated costs. GPU metrics is currently in Alpha mode, so if you are interested in trying it out, please reach out and we can enable it for you to play around with!
Cost Analytics Enhancements
We also shipped several improvements to the Costs experience so users can better analyze historical spend.
Highlights include:

A new graph that surfaces pod count and CPU/memory history in the Costs UI, making it easier to relate changes in resource requests to changes in cost over time.
Saved views in Costs that now support configurable columns and preserve those column selections, so teams can create and share consistent reporting layouts.
New difference columns (CPU request vs. limit, memory request vs. limit) in both the UI and API to quickly identify over‑requested or under‑limited workloads that may be driving unnecessary spend.
Multiple fixes to metrics and cost aggregators, downsampling lag, and numeric precision so that historical cost charts and tables respond more quickly and reflect data more accurately.

Together, these capabilities make the Costs section more useful for ongoing optimization work, beyond just monthly or quarterly reporting.
Bug Fixes and Enhancements

Fix CVE-2025-22874
Process validationadmissionpolicies in Insights
Allow API requests to reference saved view names
Add Kyverno policies to policies API
Allow users to collapse cards
Implement on demand jobs APIs
Insert validation reports
Add Options field to ListOnDemandJobs and update related tests
Support to Kyverno admissions at FE
Enhance health check endpoints
Add Kyverno admission report findings to Action Items
Add support to ManagedBy param to action items
Enhance UTMStackIntegration with TLS options and update codegen version
Add support to require SSO when calling Admin APIs
Add AffectedActionItems to responses
Improve database initialization to support optional required connections
Add support for portal ID on admin panel
Allow to add a Description to a Jira ticket
Add endpoint to retrieve all custom fields for an org
Update frontend for handling Jira custom fields
Store Kyverno policies in DB via API
Allow mapping Kyverno policies to app groups
Add endpoint to return Kyverno policies for a cluster
Implement notification process workflows and remove cronjob
Add Kyverno admission blocks to admission API
Refactor activity workers in GitHub worker to improve concurrency
Add ValidatingPolicy and remove validate from push kyverno-policies
Update policy UI to reflect Kyverno policy application status
Handle new Kyverno policy types in Insights
Add “Managed By Fairwinds” label at Policies UI
Allow mapping Kyverno policies not managed by Insights
Use new aggregators with GPU data
Insert GPU metrics into Insights
Add CPU request/memory request/memory limit/cpu limit differences

Questions About Fairwinds Insights or Kubernetes?
Please reach out to us with any questions about Fairwinds Insights or how you can manage Kubernetes at scale without putting additional strain on your internal development and operations teams. We’d be happy to walk through any questions you have about the latest capabilities in Fairwinds Insights to help you take advantage of the new functionality we’re building. Please also consider joining the Fairwinds Community on Slack as well.
If you don’t want to spend time managing Kubernetes in-house, we also provide Managed Kubernetes-as-a-Service, a people-led service that builds, delivers, and maintains mission-critical Kubernetes infrastructure.
To stay up to date with Fairwinds Insights updates, read our release notes.

*** This is a Security Bloggers Network syndicated blog from Fairwinds | Blog authored by Munib Ali. Read the original post at: https://www.fairwinds.com/blog/fairwinds-insights-release-notes-kyverno-integration-gpu-metrics

About Author

AndyC

Andy Curtis is an award-winning security consultant, researcher and public speaker. He has been working in the computer security industry since the early 1990s, having been employed by state and federal government, leading healthcare and banking providers across three continents. He has given talks about computer security for some of the world’s largest companies, worked with law enforcement agencies on investigations into hacking groups, and is a regular voice on TV and radio explaining IT security threats.

See author's posts