Using Parallel Nodes
Run multiple branches of your workflow at the same time with parallel nodes.
Overview
A Parallel node fans the workflow out into several branches that execute concurrently. Each branch runs its own sequence of nodes, and the workflow only resumes past the parallel block once every branch has finished.
Use parallel nodes when the branches do not depend on each other. They cut wall-clock time on independent connector calls, notifications, or data fetches that would otherwise sit waiting for each previous step to complete.
Configuration
- Drop a Parallel node onto the canvas.
- Set the number of branches in the node properties panel.
- Connect each branch output to the first node in that branch.
- Build each branch as a normal sequence of nodes.
How It Works
- All branches start at the same time when the parallel node is reached.
- The workflow waits at the parallel join until every branch finishes.
- If any branch fails, the workflow follows the error handling configured on that branch (retry, continue, or fail the step).
- Outputs from every branch are available to downstream nodes by step name.
Usage Examples
- Multi-source fetch - one branch pulls products from Shopify, another pulls inventory from NetSuite, and a downstream Transform node merges the two results.
- Multi-channel notify - separate branches post to Slack, send an email, and create a Monday.com item, so the slowest channel does not block the others.
Tips
- Keep branches independent. Any cross-branch reference forces ordering and defeats the point of running them concurrently.
- If branches are doing the same work over a list, use a Loop node instead. Parallel is for distinct branches, not iteration.
- Name the last node in each branch clearly so the merged step output is easy to reference later.
Common Pitfalls
- Connector rate limits hit faster when calls run in parallel. Watch for 429 errors and stagger work if needed.
- A single branch failure can stall the join. Configure retry or continue-on-error on flaky steps.
- The workflow does not preserve any specific ordering between branch outputs. Do not rely on which branch finishes first.
Related Articles
- Using Loop Nodes
- Using Condition Nodes
- Understanding Retry and Error Handling
- Adding and Connecting Nodes
- Working with Variables and Templates