State Management
State management in Orcher allows workflows to maintain persistent key-value state that survives restarts, crashes, and replays. You can store progress indicators, intermediate results, and status flags using set_state() and get_state() — and external systems can query that state in real time via registered query handlers, all without blocking the workflow.
Key Characteristics
- Durable — State is reconstructed from the journal on replay, so it survives crashes
- Scoped — Each workflow execution has its own isolated state
- Type-Safe — Compile-time type checking via serialization/deserialization (JSON)
- Queryable — External systems can query state through
registerQueryHandlerwithout blocking execution - Replay-Safe — State is consistent across first execution and replay
Workflow
set_state("status", "processing")
execute_task(work)
set_state("progress", "75%")
set_state("status", "done")
◄►
State (KV)
status"done"
progress"75%"
query
External Client
read state anytime
Explore State Management
Setting & Getting State
Learn how to set, get, and persist state values across workflow executions and replays.
State Patterns
Common patterns for progress tracking, status machines, query handlers, and state vs variables.
Best Practices & Performance
Guidelines for descriptive keys, typed state, performance optimization, and troubleshooting.