I need to trigger a dynamic number of sub-workflows in parallel (around 100)

and wait for ALL of them to complete before continuing the main workflow.

I’ve implemented a solution but I’m wondering if there’s a better or more elegant approach.

**My Current Implementation:**

**Main Workflow:**

1. Triggers N sub-workflows via HTTP Request node with `Execute Once: false`

2. Passes the total count of sub-workflows (e.g., 100) to each sub-workflow

3. Uses a Wait node (Wait for Webhook Resume) to pause execution

**Each Sub-Workflow:**

1. Performs its work (calls external LLM via MCP, takes 2-5 minutes)

2. When complete, sends an HTTP callback to a “Collector Workflow”

**Collector Workflow:**

1. Receives callbacks from completed sub-workflows

2. Maintains a counter of completed tasks

3. When counter reaches N/N (all sub-workflows finished), it resumes the main workflow via webhook

**My Questions:**

- Is this a good pattern for handling 100+ parallel sub-workflows?

- Is there a built-in n8n feature or node that can handle this more elegantly?

- Should I be using the Merge node differently, or is the callback + counter approach appropriate?

- Are there any performance considerations or limitations I should be aware of?

**Key Requirements:**

- Dynamic number of sub-workflows (not fixed)

- Each sub-workflow is long-running (2-5 minutes) use callback solution

- Main workflow must wait for ALL sub-workflows to complete

- Need to collect and process all results together

I’m open to completely different approaches if there’s a better way to architect this!

**Main Workflow Structure:**

**Sub-Workflow Structure:**

when callback return from the long process → call to the Collector workflow

**Collector Workflow Structure:**

The final output is a collection of all sub-workflow results that need to be merged and processed together in the main workflow.