10.2 Node Lifecycle

Go To Top

Return to Node Implementation Pattern

Each Phase 0 Node should follow a defined Lifecycle. The Lifecycle gives the team a shared way to describe Node execution, observe Node state, and diagnose runtime behavior.

The Node Lifecycle should support the following operational states, where applicable:

1. Starting
2. Running
3. Degraded
4. Recovering
5. Stopping
6. Stopped
7. Failed
8. Unknown

A Node should publish Lifecycle state changes through the approved Control Plane status topic. The Node should report Starting when it begins initialization, Running when it completes startup and can perform its role, Stopping when it begins controlled shutdown, Stopped when it completes controlled shutdown, and Failed when it cannot continue safely.

A Node may report Degraded when it can continue running with reduced capability. A Node may report Recovering when it attempts to restore normal operation after a recoverable failure. The team should define clear criteria before using Degraded or Recovering so that those states carry consistent meaning across Nodes.

The Unknown state should describe an observer's lack of current status, not a normal state that a healthy Node publishes about itself.