Distributed Nodes
Status: Planned — This feature is under active design and not yet implemented.
Overview
Distributed Nodes will allow armament sessions to span multiple machines. A leader node coordinates work distribution, while worker nodes execute agent tasks with automatic failover and session checkpointing.
Planned Capabilities
- Leader election with automatic failover
- Worker pool registration and health monitoring
- Cross-machine session handoff
- Distributed task queuing with priority scheduling
- Session checkpoints for crash recovery
- Network partition handling with split-brain prevention
- Cluster-wide budget enforcement
- Secure inter-node communication (mTLS)
Related Future Work
- Cluster UI — Visual dashboard for cluster state
- Graph Workflows — DAG execution across distributed nodes