Remote Screen Control (Beta)
Node detail · Beta
Handle problems you cannot solve over the phone by looking at the screen yourself. Request a connection from the manager and the screen is streamed as-is, with mouse and keyboard control. PCs behind a router that cannot connect directly are relayed through the server, and the user of the remote PC sees a persistent notice for the whole session.
Team Lead Jung — AI Diagnostic Assistant (Beta)
Dashboard widget · Beta
No more opening several screens to work out why a PC is slow. Ask in plain language and it inspects PC status, processes, and collected data on its own, then answers and shows the evidence behind its judgement. It can diagnose from server-side data even when the PC is powered off.
PC Asset Inventory
Node management widget
Hardware and installed-software details for every PC load in one pass. Previously only 300 machines were shown; that cap is gone. Start from group cards built by org chart, IP range, OS, or manufacturer, and apply requirements such as GPU or memory to surface the groups that meet them.
Server Health Diagnosis
Server monitor widget
A Normal / Caution / Needs attention verdict now appears at the top of the server status screen together with the reasons behind it. How fast storage is filling, when it will run out at that rate, and which jobs are currently running long are all on the same screen. Previously anything that was merely connected counted as Normal.
Update Agents Straight from the PC List
Node list · Actions
Select one or many PCs and run Agent Update; each PC compares its own version against the latest and downloads only when they differ. Builds marked unstable are skipped, and mass runs are staggered so downloads do not all land at once.
Load Index Scoring Reworked
Values may look different
The cost of a program repeatedly starting and stopping is now converted from actual measurements into the score. PCs with many short-lived windows no longer score higher than they should. Rankings may look different from before — this is the scoring basis being corrected, not a fault.
Commands Stuck In Progress or Showing "No Response"
Command management
Failures that would repeat identically (timeouts, permission denied) now stop retrying and are marked finished. Previously they were resent every 15 minutes for up to seven days, and each resend blocked that PC from taking other commands for a minute. Commands that had actually run on a restarted agent but showed only "No response" are fixed as well.
Symptoms Silently Dropped Under Load
Values may look different
The way symptoms are recorded has changed so nothing is lost when they arrive in bursts. Previously anything beyond processing capacity disappeared with no indication that it had. Counts for busy periods may therefore be higher than before.
Live View Dropping in Large Fleets
Large deployments
Fixed the live view dropping every eight minutes and newly powered-on PCs failing to appear in environments with many connected machines. The number of simultaneous connections the server sustains has been raised substantially. The symptom reproduced at around 1,000 machines.
Connection Blocks Told Apart by Cause
Detection rule
When a program cannot reach a server, cases where the PC's own firewall blocked it are now reported as Local firewall block instead of being lumped under Access denied. You can tell immediately that it is not a server or network problem.