Lifecycle and Monitoring¶

Studio lets you manage runtime processes from the canvas without leaving the graph context.

Service Toolbar States¶

Each service node has a compact process toolbar for disable/start, activate/stop, sync, and restart actions.

Status	Screenshot	Meaning
Not Run		Process not started
Disabled		Excluded from compile, deploy, and auto-start
Running		Process is alive
Paused		Process is alive but inactive

Deploy Flow¶

1. Edit the graph on canvas
2. Validate required state fields and host-service bindings
3. Compile the runtime graph
4. Send the compiled rungraph to target services
5. Watch service monitor rows and logs for acceptance or rejection

Monitoring Views¶

Service Manager is the central dashboard for runtime health. Watch these first:

• CPU/RAM/GPU usage spikes
• latency and tick drift
• error counters and repeated failures
• alive/ready/active transitions after deploy

Good Operational Habits¶

• Start with infrastructure services first, then hosts such as f8.pyengine, then downstream producers/consumers.
• Restart only the service that owns the failure when possible.
• Keep Disabled nodes in the graph for experiments, but avoid leaving them wired into production scenarios without notes.