Right-sizing recommendations can now be customized to give teams more flexibility in how they act on RealTheory’s guidance. This allows you to adjust CPU and memory configurations to better match operational needs while still ensuring baseline compute requirements are met.
Right-sizing recommendations now offer an enhanced details panel that displays CPU and memory charts for both current and recommended resource requests. The panel also highlights whether the recommendation will provide cost, performance, or reliability optimizations (or a combination of these benefits).
For users who prefer the previous layout, the legacy details panel remains available and can be accessed at any time via a simple toggle switch.
KIP, RealTheory’s conversational AI assistant, is now available within Reports, enabling you to quickly query and interpret cost, usage, and performance data without leaving the view. KIP can answer questions, explain trends, and guide you to relevant details, helping you analyze report data more efficiently.
We’re pleased to introduce KIP, RealTheory’s conversational AI assistant, designed to help you access insights more efficiently. Using natural language queries, KIP can help you interpret recommendations, uncover optimization opportunities, and quickly retrieve relevant data to enhance your experience and improve productivity.
Right-sizing recommendations are customized based on the optimization goal you set, providing more relevant guidance for adjusting CPU and memory resources.
These recommendations reflect each resource's active Optimization Profile, which prioritizes performance, cost savings, or a balanced approach. Profiles can be assigned manually or applied at scale using Optimization Policies, allowing flexible alignment with your organizational structure.
Users can now control whether newly deployed agents start as active or inactive and can update the active/inactive state of agents already running in their environment. This provides teams with greater flexibility in managing agent behavior and data collection.
Previously, object name links always opened the source object in the Visualization view. You can now set a preferred view—either Visualizations or Quick Views—on your My Profile page, and object links will default to that selection.
A new menu next to each link lets you choose between both views, making it easy to select the most appropriate option without changing your default.
You can now specify the Impact level (High, Medium, Low) when creating a custom version check rule, making it easier to indicate how critical it is when an application is running a non-compliant version.
Resolved an issue where, in environments with a large number of Groups, the Find bar in the Add Groups section of the Add/Edit User modal was only searching a subset of Groups. This could cause valid search matches to be missed.
Resolved an issue where sorting and searching on the Settings > Team > Users view was case sensitive. This caused searches to return only exact-case matches and sorted results to be split into separate upper- and lower-case groupings. Sorting and searching are now case-insensitive.
Users can now select Visualizations, Quick Views, Recommendations, or Alerts as their preferred landing page when logging in to RealTheory. In addition, clicking the RealTheory logo in the Console will quickly take you to your preferred view. Configure your preference in the My Profile view.
RealTheory SSO now supports processing claims from identity providers to automatically assign users to groups and roles based on identity provider metadata. This includes support for creating groups from claims, assigning users to groups, and assigning RealTheory roles to users.
Added support for Google Identity Platform as a single sign-on (SSO) identity provider. When configuring SSO for the Google Identity Platform in the Console, the email OpenID Connect (OIDC) Scope must be selected.
Users can now view logs for any container in a pod by selecting the preferred container from the Containers filter option on the Logs tab of the pod's information panel. Note: If the pod is running only one container, the filter option is not available.
The RealTheory agent, the Collector, can now be restarted or upgraded from Settings > Agent > Management. In addition, the Collector logs can be accessed from this view.
Data in the cluster Quick View can now be downloaded in .csv format for export to third-party applications.
Filter options for the Total Cost, Lost Dollars, and Utilization reports can now be quickly reset to the default values.
Users can now view the previous log for any container in a pod by selecting the Show Previous filter option on the Logs tab of the pod's information panel. The previous log frequently contains helpful troubleshooting information.
Double-clicking on a row in the Total Cost, Lost Dollars, or Utilization reports will drill-in to the data for a subordinate object type of the parent object. Use the dropdown arrows in the breadcrumb to select a different subordinate object type.
Double-clicking on a row in Quick View will drill-in to the data for a subordinate object type of the parent object. Use the dropdown arrows in the breadcrumb to select a different subordinate object type.
RealTheory now scans public container images in your cluster for vulnerabilities automatically. To scan private images, register registry credentials in Settings > Security > Container Vulnerability Scanning. View results in the Vulnerabilities tab of the pod information panel.
The Node Pool tab on the information panel for a cluster now includes CPU and memory utilization information for each node pool to assist in making right-sizing decisions.
Users can now filter recommendations by object type on the Recommendations view and the Recommendations tab on the information panel. The Object filter complements existing filters to provide more granular filtering.
Users can now filter alerts by object type on the Alerts view and the Alerts tab on the information panel. The Object filter complements existing filters to provide more granular filtering.
To optimize Collector performance on larger clusters, you can now set the Approximate cluster vCPU count in Settings > Agent > Deployment > More Options. This ensures appropriate CPU and memory requests are applied to the realtheorycollector pod.
RealTheory now fully supports Kubernetes clusters with ARM64 architecture nodes, including analysis, cost reporting, and optimization insights. Set the appropriate architecture for the RealTheory Collector in the More Options section of Settings > Agent > Deployment.
On This Page
Search for a command to run...