A Redesigned Agent Details View, and Updates That Don't Wait on Offline Hosts
September 26, 2026 · InfraScout Team
This release reworks the view you open for a single agent, and fixes two fleet behaviors that made a healthy fleet look broken: agent updates that stalled behind offline hosts, and commands recorded as timeouts that had never reached their host.
Agent Details in two panes
The Agent Details view used to be a stack of collapsible sections with the actions in a footer. It is now two panes.
The left rail answers the first questions you have about a host: its name, which you can rename in place, its status, its hostname and its agent ID, with a copy button so you no longer select a long identifier by hand. Below that sit when it was last seen, whether it is ready for commands, its version — with a badge naming the newer version when an update is available — and how many days its certificate has left. Force Renewal, Revoke and Delete are pinned at the bottom of the rail.
The right pane holds the detail. System separates the operating system's name from its build and shows the architecture, the account the agent runs as, and when it registered. Capabilities lists what the agent can do, and Certificate draws the certificate's validity as a timeline from issue to expiry.
The view also reflects your changes straight away. A rename or a forced renewal used to show its success message while the view kept the old data until you reopened it; the view now reloads the agent after each change.
Offline hosts no longer hold up agent updates
Agent updates go out a limited number at a time, oldest first. When the oldest rows in the queue belonged to hosts that were offline, they took up every slot on every pass, so nothing behind them was sent — and the Updates board then painted the entire online queue as Late. A handful of decommissioned laptops could hold back an update for the whole fleet.
Rows for offline or deleted hosts are now passed over, so reachable hosts keep updating while the offline ones wait until they reconnect.
Commands no longer time out while an agent reconnects
Agents briefly reconnect during our own updates. A command sent at exactly that moment could miss the agent, wait out its entire timeout plus half a minute, and then be recorded as a timeout, even though it had never reached the host. Now the command follows the agent to its new connection and runs. If the agent cannot be reached at all, the command fails promptly with an error that says so, rather than looking like a host that hung.
More detail: Agents (Admin) — Agent details and Agent Updates.
Questions or feedback? Reach us at info@infrascout.cloud.