Flow Studio
Pause and Resume
Pausing stops the workflow at the next checkpoint — after the currently running node completes. The execution remains in a suspended state on the server, preserving all execution memory, until you resume or cancel it.
How Pause Works
Click the Pause button in the Top Toolbar during a running execution. What happens:
- A pause signal is sent to the backend:
POST /api/v1/process-engine/executions/{id}/pause - The currently running node finishes (pause does not interrupt mid-execution)
- Before picking up the next node, the Process Engine checks the pause signal and suspends
- The execution status changes to
Paused - The Observer Panel shows the Paused state with elapsed time frozen
Pause Is Not Immediate
If the current node is slow (e.g., a 10-second HTTP call), the pause takes effect after that node finishes. You cannot pause a node mid-execution. To abort immediately, use Cancel instead.
What Is Preserved During Pause
- All completed node outputs (ExecutionMemory is fully preserved)
- All workflow variables set so far
- The next node to be executed (cursor position in the graph)
- Execution log entries from completed nodes
- Elapsed time (shown on resume)
Resuming
Click Resume to continue execution from the suspension point. The Process Engine continues with the next queued node. Resume sends: POST /api/v1/process-engine/executions/{id}/resume
Pause Timeout
A paused execution does not wait indefinitely. If not resumed within the configured timeout (default: 24 hours), the execution automatically transitions to Failed with the reason "Pause timeout exceeded". This prevents orphaned executions from consuming resources.