Migration of VMs between CRs is limited by migrationtimeoutminutes which has a "hidden" default of 180 minutes, denoted by "migrationtimeoutminutes": 0 in /etc/solus/agent.json

Increasing the timeout (e.g. to allow successful migration of large servers) requires manually editing the agent.json on source and destination CRs.

However, if you run VMs large enough that it makes sense to increase the timeout, you will most likely need to increase this across all CRs - otherwise you'll hit intermittent failures as you migrate VMs around over time.

It should be possible to manage the timeout centrally from the management node, so that all newly provisioned CRs also have the increased timeout set by default.

It may be desirable to extend this functionality to cover central management of agent.json in general, but at the moment we have only had a need to modify this specific parameter.

The general idea is that CRs should be managed as a group, to avoid per CR configuration drift.