In an Automation Suite production environment, server nodes host the Kubernetes control plane and run the distributed etcd datastore. Maintaining a functional cluster requires an active etcd quorum, which follows the Raft consensus algorithm where a strict majority of voting members ($(N/2) + 1$) must remain online and communicating. In a standard three-server HA architecture, quorum requires at least two functional server nodes. If two server nodes fail simultaneously, only one server remains, breaking quorum and causing the control plane to transition into a read-only or completely halted state. Consequently, the entire cluster becomes unavailable: API requests fail, pod scheduling halts, and the cluster cannot reconcile state. While agent nodes may continue running existing compute tasks temporarily, losing control-plane consensus renders the cluster non-operational until quorum is restored. (Reference: UiPath Automation Suite High Availability Design – Control Plane Quorum and Failure Scenarios).
===============
Contribute your Thoughts:
Chosen Answer:
This is a voting comment (?). You can switch to a simple comment. It is better to Upvote an existing comment if you don't have anything to add.
Submit