Why Does the Local Agent Automatically Exit After Being Idle?
Understand the exit mechanism after about one hour of no command activity and how to reconnect as needed.
Idle Exit Is a Client Behavior
Currently, the Local Agent will automatically exit after approximately one hour without receiving any command activity to reduce resource usage. The terminal will display messages indicating idle timeout and automatic termination. This is different from issues like account subscription expiration or password changes, and you cannot assume the CLI remains online just because you can still log in via the web. The desktop application has its own connection mechanism and should not directly apply this CLI timing rule.
Confirm It Is Idle, Not Another Fault
Check the last few lines of the terminal running the client: idle exit, token rejection, command relay timeout, and connection limit each have different prompts. Do not immediately reset the token just because the machine disappears from the list. If the connection is lost during a long task, first verify the original task and the actual status on the local machine; do not assume the execution is complete or safely terminated simply because there is no output in the window.
Re-verify the Target Machine When Recovering
Confirm that only the client has exited, then copy the command again in Settings → Remote Control and start it on the target computer. Wait until 'Local sandbox is ready!' appears. Confirm the hostname and environment selection on the web page before continuing with the remaining work. Do not send repeated execution requests without actual purpose just to stay online; if you need to keep a task available for a long time, also monitor the target machine’s network, sleep, and client status.