When using Nerdio autoscale feature with a minimum number of session hosts configured, the minimum hosts remain powered on continuously and do not participate in scale-in operations. This results in extended uptime (e.g. 20+ days), which can lead to performance issues such as high CPU utilisation and degraded user experience. Currently, there is no way to target only these long-running hosts for reboot.
We require a capability to identify and selectively reboot session hosts based on conditions such as uptime or performance. The existing option to schedule a reboot applies to all hosts in the host pool, which is not suitable in environments where only a subset of hosts (e.g. minimum hosts) require cycling.
Introduce a feature within Nerdio Manager for Enterprise that allows conditional reboot logic for session hosts.
This could include:
-Filtering hosts based on uptime (e.g. >7 days)
- Checking session state (e.g. no active users or drain mode)
- Applying reboot actions only to selected hosts instead of the entire pool
Ideally, this would be integrated into autoscale policies or available as a scheduled action within Nerdio. This would help maintain host performance without impacting all session hosts or requiring custom runbooks.
Comments (0 comments)