Server Settings.

Manage the settings of individual servers. Server settings control the type of instances that can be deployed to each server as well as resource allocation limits.

Basics.

In the basics section of server settings users can find more information about the server Id and hostname.

Update Server.

Nickname

Servers can now be nicknamed, giving the user control over naming schemes. These changes do not impact the underlying hostname of the server.

Tags

Server tags are a way to give instances more information for deployment. Cycle will automatically generate a few relevant tags.

Click into the Tags field and start typing the tag to add. When ready, press enter. To remove a tag, click the [x] next to it.

Allow Pool

If this checkbox is selected, this server will accept instances that have no tags specified.

Allow Services

When an environment is created, the service containers that come with that environment are automatically created and placed on servers. If this box is selected, the server will accept service containers. If it is unchecked, no service containers will be deployed to this server.

Advanced Updates.

Allow Overcommit

To allow for the server to over-commit its CPU resources, check the box and enter a multiple. The multiple is how many times over-committed the CPU can be.

Controls.

Controls is split into two sections:

1. 01
   General - where the compute service can be restarted.
2. 02
   Advanced - where the server or compute spawner can be restarted.

Restart Compute

Restarting compute will simply restart the compute process running on the node. This restart will not interact with networks or containers running on the node. Minimal to zero downtime impact.

Restart Compute Spawner

Restarting compute spawner will restart all containers on the node to restart and will rebuild all networks from 0 to 100%. This restart will cause downtime for all container instances on the node.

Restarting the Server

The platform offers two mechanics for restarting a server. Hard restart and soft restart.

• Hard Restart - uses the providers API to provide the restart.
• Soft Restart - uses software mechanics to reboot the server.

Both of these options will allow for the server to update if needed.

If the server is at Google Cloud Platform (GCP) - use Soft Restart unless a hard restart is absolutely needed. GCP's platform, as of this writing in 2026, will actually release the underlying VM resource and re-reserve it which can result in possible out of inventory issues where the VM that was previously reserved cannot be renewed.

Due to the nature of soft restart being software/command based, it may not be enough to use soft restart if a machine with 2 cores has a load of 8,000 - or if other resources are so exhausted that commands are failing to get through to the scheduler.

Evacuate Server.

Users that wish to move all workloads off of a server and onto another can use the evacuate server form to select a server target to migrate the server instances to.

Deleting a Server.

Use the delete from to delete the server.

The "Force Delete Server" checkbox is to be used only when a normal delete action cannot be done and the user needs to remove the server forcibly from Cycle.