Ping check
Node list · right-click action
Right-click a node, a group or an entry in the detail view and the server sends a ping to that PC's address. The result comes back as reachable, no response, or cannot determine, together with the address it was sent to. A node showing as online means the agent reached the server; ping travels the other way, so an online PC may still be unreachable. Allow incoming ping on the endpoint firewall - if it is blocked, a healthy PC also reports no response.
Team Lead Jung inside the command and symptom screens
Command management · Symptom detail
Previously you could only ask from the Team Lead Jung tab in the node detail view. Now you talk to it right inside the command screen and the symptom detail, while looking at them. When writing a command, describe the work in plain sentences and it drafts the command and fills the form for you. When reading results, it judges whether they answer the original request and arranges them in the format you want. The symptom detail gained a Team Lead Jung tab that investigates the cause, the place it happened and whether the file involved is safe, and reports a conclusion.
Whole web services watched as one
Server monitor
Register a web service once and name resolution, connection, certificate and response are all checked in one pass, telling you which step broke. Targets with a certificate nearing expiry are flagged as a warning rather than an outage (30 days ahead by default, adjustable per target). Targets you already registered are migrated automatically, so there is nothing to register again.
Symptom summary on the home dashboard
Home dashboard
A symptom summary you can place on the home dashboard is new, and the home figure now matches the symptom screen. The two screens used to count differently, so the same period showed different numbers. The home summary excludes routine, recurring symptoms by default.
Faster agent start and stop
Agent
Time from boot to appearing in Manager fell from 22.8 to 4.6 seconds, and shutdown from 64 to 6.9 seconds. The health check interval moved from 10 minutes to 1 hour, and hardware details that were re-collected every 5 minutes while the server was unreachable are now collected only when the server address changes. Everyday resource use on the PC dropped as well.
PCs shown online or offline incorrectly
Agent
Four causes were resolved: Wi-Fi PCs flipping state repeatedly, PCs failing to connect when their clock differs greatly from real time, a damaged settings file leaving a PC permanently offline, and settings reverting on every reboot on PCs that use recovery software. Existing customers keep the previous behaviour.
Faster first answer and wider checks in AI diagnosis
Team Lead Jung
The first answer in a new conversation dropped from 60-72 seconds to 18-29 seconds. Open questions such as "is anything wrong?" now check hardware, security updates, memory and storage without skipping any, and give a verdict per item. Caps on web search count and results are gone, citations open the source directly, and work that requires a reboot is followed up: once the PC comes back online the check runs and the result is reported.
Session lifetime and account control
Settings › Accounts
You are signed out after 12 hours of inactivity, and after 24 hours from sign-in even while working. Sessions previously never expired. A disabled account is refused at sign-in and any open screen is cut off immediately. A super administrator can change another account's login ID, so there is no need to delete and recreate the account.
Monitoring checks falling behind
Server monitor
The number of targets checked at once rose from 20 to 60 and can be tuned to the size of the site. Unresponsive devices no longer push healthy ones out of turn. A device that fails three times in a row has its interval doubled up to eight times, but never beyond 30 minutes.
Port list you could not act on
Server monitor
Each port now shows the description the administrator wrote on the device, so you can tell which port it is at a glance. Errors are shown as the increase since the last check instead of a running total. Previously only the total accumulated since the device booted was shown, so there was no way to tell 300 days of history from a problem that started this morning.
DNS failures raised as symptoms
Symptoms · Rules
Three situations that went undetected are now caught: the DNS server not responding at all, responding with an error code, and the lookup failing because the network is down. The record includes which program was looking up which domain, and the address of the DNS server that did not answer. Closed networks with external DNS blocked, or an internal DNS that has stopped, fall into this category.
1:N commands built from the PCs you selected
Command management
Select targets in the node list or the org chart and press Create command: the command management screen opens with those targets already set, rather than a separate window. Your selection is carried over so you can continue writing.
Load figures left behind and departments lost
Server
Deleting a PC now clears its hourly load aggregates as well. They used to remain, so querying a program used only on that PC raised an error. A failed org-chart lookup was also indistinguishable from "no department", which let a temporary failure overwrite and erase department information; that is fixed.
Wrong relay address accepted without warning
Install · Update
Remote control reaches PCs behind a router by relaying through the server. If the external address used for that was wrong, the connection simply failed with no reason given. Install and update now ask for that address with examples for each situation. The server certificate expiry is also capped at the issuing authority's expiry.