中文
Control and State Nodes

Control and State Nodes

Entry, branching, loops, Run state, time, and randomness

Frequently used control nodes

Node Use Caution
Run Started Main Run entry Emits once per Run
Branch Choose true or false Calculate the condition upstream
Delay Wait before continuing Prefer observable conditions when available
Repeat Execute a bounded region Add interval and exit behavior
For Each Process a typed list One body activation completes before the next
Retry Region Retry routed failure Retry transient failure only
Strongly Typed Switch Choose among cases Clearer than deeply nested branches

Continue and Break apply to their owning region, not an arbitrary outer loop.

Run state

Read State, Write State, Add State, State Metadata, and State Last Change operate on typed state initialized separately for each Run. Use state for counters, flags, and recent values inside one Run; use an explicit durable destination for data that must survive application restart.

Time and randomness

Observe Time and Stopwatch nodes measure workflow intervals. Random Integer, Number, Boolean, and Choice produce values recorded as Run evidence, rather than pretending to be deterministic math.