AI / Insights
Features
- Claude Opus 4.6 model support: Added support for the new Claude Opus 4.6 model via both Anthropic and Vertex AI providers, expanding the available AI models.
- AI Investigation notifications: AI Advisor investigation results are now automatically delivered through configured notification channels (Email, Slack, Jira, ServiceNow) with proper formatting for each channel.
- AI Advisor message manager: Implemented intelligent token management for AI Advisor conversations, enabling longer and more productive sessions without hitting token limits.
- AI Advisor audit log entries: AI Advisor requests now appear properly categorized in the Audit Log with session names and IDs, and payload details can be viewed without errors.
- KMI agent discovery: The KMI agent is now discoverable through the AI Gateway agent discovery endpoint, enabling integration with AI Advisor workflows.
- Improved duplicate alarm detection: Alarms with the same rule but different dimensions are now treated as distinct issues, allowing separate AI investigations for different alert contexts.
Bug Fixes
- AI Advisor crash prevention: Fixed an issue where the AI Advisor page could crash when the LLM service returned a null or empty response.
- AI Advisor resize handle overlap: Fixed a UI issue where the split window resize handle in AI Advisor overlaid the navigation panel.
- Device config tool toggle clarity: Improved wording on the AI Advisor Command Access settings page to clarify what enabling configuration access does.
Alerting
Features
- CFMT BGP Routes API migration: CFMT mitigations now use the BGP Routes API for advertising and withdrawing BGP routes, with optimized fetching and a feature flag for gradual rollout.
- Mitigation platform health status: Platform health reporting now uses a descriptive health status (HEALTHY, WARNING, FAILING) instead of a simple boolean, with improved error reporting and visibility.
- Custom notification templates for AI investigations: Added AI Investigation event types to custom notification templates, enabling customers to receive formatted notifications for AI investigation events.
Bug Fixes
- Alert pagination fix: Fixed an issue where alert result lists were truncated to 10 results instead of showing the full result set.
- Policy dimension grouping persistence: Fixed a bug where the "Group Dimensions" toggle in alert policies was not persisting after being enabled.
- Shadow policy baseline loading: Fixed a bug where shadow policies were not included in the baseline loader, ensuring complete baseline data collection.
- Default Universal Agent alert fix: Fixed the default Universal Agent alert policy which was not functioning correctly.
Cloud
Features
- OCI DRG route tables in Map: Route tables are now displayed on the OCI entity sidebar in the Map view for DRG resources.
- AWS VPC/Subnet lookup in Data Explorer: Added efficient lookup summaries for AWS VPCs and subnets, enabling faster filtering by VPC Name and Subnet Name.
- CloudPak license page metrics: Updated how cloud metrics are calculated and displayed on the license management page for more accurate usage data.
Bug Fixes
- CloudPak license card fixes: Addressed UI issues with the CloudPak card on the license page, including dark mode color corrections.
- FlowPak plan query accuracy: Corrected inaccurate flow rate values on the License Page for FlowPak plans that were showing inflated values due to improper aggregation.
Core
Features
- Flow sample rate indicators: Added comprehensive flow sampling visibility to the Device Details page, including sample rate status indicators (Downsampled/Low Resolution), 30-day traffic charts, and actual vs. configured sample rates.
- Device downsampling filters: Users can now filter devices by downsampling status (None, Low, High) in the device settings view with warning tooltips.
- Audit log enhancements: Added IP address and user agent fields to the Audit Log, and replaced the Audit Log table with an improved search interface.
- GetDevice API expanded fields: The GetDevice API now returns device OS name, OS version, and serial number.
- Device site country enrichment: Metric measurements are now enriched with device site country data, allowing customers to filter guided mode dashboards by device location.
- Additive dictionary permissions in UDE: Implemented additive dictionary permissions for the Unified Data Explorer for more flexible permission management.
- Unit filters in UDE: Added UI controls for unit filters (DNS codes, HTTPS status codes, health metrics) and search capabilities for synthetic agents and tests.
- Enumerations API for tagging: New gRPC-based enumerations service for faster tag lookups with reduced query times and database bloat.
- DNS Analytics collection option: Customers can now enable DNS Analytics collection in kdns configuration to gather additional fields from DNS response headers for deeper analytics.
- LiveSynth connectivity testing: Enabled new infrastructure for LiveSynth connectivity testing services with gRPC and SSE protocol support.
- Compatible policy validation: The system now automatically determines and displays only compatible policies when adding policies to a tenant, eliminating trial-and-error.
- Flexible metrics plan support: Kmetrics-ingest now accepts either metrics or npak plan types during the transition period.
- Nokia SR-OS Classic device support: Added support for Nokia SR-OS Classic devices in ksqueegee.
- UA capability config grouping: Configuration options can now be organized into logical groups with feature flag support for conditional display.
- Device inventory performance: Device metric sparkline data is now loaded via a single backend API call instead of multiple frontend queries, improving page load performance.
Bug Fixes
- IS-IS tab 500 error fix: Resolved a bug where the IS-IS tab on Device Details pages returned a 500 error, preventing customers from viewing IS-IS adjacency information.
- Resizable columns improvements: Fixed issues where resizing table columns could push action buttons off-screen, cause text oscillation on Safari, or misalign columns.
- ISIS relations save error: Fixed an error that occurred when saving device changes with ISIS adjacencies.
- Library page crash fix: Added safety checks to prevent crashes when using grouping and filtering together on the Library page.
- Deleted devices on dashboards: Improved cleanup of deleted device references from dashboards.
- Device flow rate display fix: Fixed a bug where Device Details failed to display flow rate measurements when device data was not fully loaded.
- Custom dimension values fix: Fixed missing values for custom dimensions in query results caused by tagging lookup caching issues.
- Tag lookup race condition fix: Eliminated a race condition in tag lookups where concurrent requests could miss values.
- Tag batch update performance: Improved performance for high-volume DNS hostname tag updates by implementing deferred batch processing.
- Query reliability improvements: Increased socket backlog capacity and implemented adaptive concurrency tuning to improve query reliability under high load.
- Vestigial query filter removal: Removed an outdated filter clause that was preventing device name filtering from returning results.
- Flow query error prevention: Added graceful fallback when proxy measurements are unavailable, preventing errors by falling back to the standard /flow endpoint.
- Exporter crash fix: Fixed a crash that occurred when stopping or restarting exporters, enabling seamless configuration changes.
- Storage rebalancing accuracy: Improved shard distribution decisions by using server-side metrics with proper data smoothing instead of client-side metrics.
- Unnecessary service restart prevention: Eliminated unnecessary ingest service restarts during storage sharding operations that were causing data loss.
- Destination limit enforcement: Added enforcement of maximum limits on alerting and external destinations per device to prevent misconfiguration.
Network Monitoring System (NMS)
Features
- SNMP interface metrics configuration: Customers with the "UA SNMP Polling for Flow Enrichment" feature flag can now automatically set the proper configuration when choosing to collect SNMP interface data via NMS.
- NetApp interface support: Interface information is now available for NetApp devices in the portal via the legacy metadata poller.
- Cisco Products MIB support: Added Cisco Products MIB to the MIB database for expanded Cisco equipment monitoring.
- Updated Infoblox MIBs: Updated Infoblox MIBs with newer versions for more accurate device monitoring.
- MIB browser improvements: Reorganized the MIB browser home page with vendor-based categorization, enhanced metadata on MIB cards, and improved loading experience.
- UA capability label rename: Renamed the "NMS" Universal Agent capability label to "SNMP/ST" for clarity.
- Agent capability state reporting: Agents now report proper capability state information instead of just OK status.
Bug Fixes
- NMS status indicator accuracy: Flow-only devices no longer incorrectly show NMS status indicators, reducing confusion about monitoring coverage.
- Bulk device deletion protection: Devices with active manually-defined connections are now blocked from bulk deletion with a clear error message.
- MIB OID highlighting fix: Fixed incorrect OID highlighting when viewing multiple OIDs in the MIB browser's full tree view.
- Notification channel persistence: Fixed an issue where notification channels were not persisting when created on subpolicy forms.
- Package page access: Package pages can now be accessed directly via URL without requiring navigation through the table first.
Synthetics
Features
- Site-to-site traffic filters: Added new filter options for site-to-site tests, enabling filtering by site and ultimate exit parameters in Subtest Test Results.
- Synthetics toggle clarity: Improved UI clarity for Synthetics toggle controls by moving the disable warning label to the switch component.
Bug Fixes
- DNS test period sync: Fixed divergent test/ping/trace periods in DNS synthetic test configurations.
- Batched flow export optimization: Optimized flow data export for large synthetic tests by batching results to prevent excessive flow volume.