Plixer ML Engine changelogs

Plixer ML Engine changelogs#

Changelog entries are displayed in the format DESCRIPTION (Ticket Number).

Note

Plixer ML Engine v19.8.0 - July 2026#

Changelog

New features

  • Auto-tune deployment based on system resources

  • Create script to backup and restore system configuration

  • Do not build models for excluded assets

  • If replica counts in the UI are set to greater than 6, increase kafka partition count to match

  • Improve messaging so user knows what setting to change to resolve threshold violations

  • Remove TrainingData-old directory to save disk space

  • Require at least a day of traffic to build a model (don’t build models for hosts that don’t have enough traffic)

  • Support high throughput mode to increase the number of assets that can be monitored

  • Support standalone deployments in AWS using an AMI

Fixes

  • Addressed various security issues

  • Exclusions are only applied to violators, not to targets (1043)

  • Make sure internal Kubernetes IP addresses don’t conflict with LAN ip addresses (1041)

  • Make sure kafka producers are reset when running a –reset (1040)

  • Offline deployments have multiple .log files in the plixer user home directory (1046)

  • Remove unused swap file (1056)

  • Rogue LDAP Service is categorized as DHCP Spoofing (1020)

Plixer ML Engine v19.7.0 - November 2025#

Changelog

New features

  • 19.5.0 ML engine HyperV image requires a gen1 VM but docs state v2

  • Add SNMPD to ML Repo

  • Support for IPv6 subnets in ML Engine rules

Enhancements

  • Configure which IP Groups Deep Graph Learning is monitoring

Fixes

  • Addressed various security issues

  • engine.lock file persists where the engine pod restarts before all the models are moved (1025)

  • Upgrade script does not remove hashicorp.list from /etc/apt/sources.d/ (1003)

  • Vitals model count report is reporting fewer models than are built (1026)

Plixer ML Engine v19.5.0 - March 2025#

Changelog

New features

  • Detect brute force activity using failed SMB logon attempts

  • Detect remote ransomware attacks using SMB read/write data

  • Support ability to modify AI Engine deployment settings from the user interface

  • Support custom thresholds for data accumulation detections

  • Support KVM

  • Support offline vSphere cluster deployments

Enhancements

  • Optimize processing to support larger workloads

  • Prompt user for deployment information rather than use configuration files

  • Support ability to modify AI engine application settings from the user interface

  • Support ability to modify AI Engine seasonality settings from the user interface

  • Support child groups for IP Group exclusions

  • Support extending expiration date of ssl certificate used by Kubernetes cluster

  • Verify application pods are updated during an online system update

Fixes

  • Addressed various security issues

  • Define violator and target correctly for brute force events (889)

  • Update rogue DHCP detection logic (954)

Plixer ML Engine v19.4.0 - August 2024#

Changelog

New features

  • Added LDAP rogue service detection

  • Added new alert type in Scrutinizer for Suricata TLS alerts

  • Added support for Hyper-V deployments

  • Added support for IP groups in exclusions

  • Encrypted Traffic Analytics (ETA)

  • New base operating system: Ubuntu

  • User behavior analytics for O365/Azure AD

Fixes

  • Addressed various security issues

  • An ML engine can now pair with a multi-collector distributed Scrutinizer (757)

  • Behavior tab missing after upgrading to 19.3 (745)

  • Certain utilities not functioning properly with 19.3 (786)

  • Improved ML event messages by including additional details (810)

  • Time zone set as UTC causing training data to not be associated correctly with workdays/weeknights/weekends (750)