Problems with Multi-User Parallel Workflows in n8n (Outputs Getting Mixed)

Questions
,
1

Hi everyone,
I’d like to ask about using one n8n instance with multiple users/workflows running at the same time.

My team is subscribed to n8n Pro.
I created one workflow that generates images. When I tested it alone, everything worked perfectly.

After that, I duplicated the workflow for 4 different team members (one workflow per person).

Problems we’re facing:

  • Output images sometimes get mixed up and are sent to the wrong team member, especially when workflows run at the same time

  • When all workflows are used together, the runtime becomes slower

I already tried adding unique IDs / codes per team member inside each workflow, but it didn’t fully solve the issue.

This makes me wonder:

  • Is n8n suitable for this kind of multi-user, parallel workflow setup?

  • Are there best practices or architecture suggestions to avoid mixed outputs and performance slowdown?

Any advice or guidance would be greatly appreciated.
Thanks in advance!

2

Hi @woncho !

This happens because multiple parallel executions are writing to and reading from the same shared external resources, so they overwrite each other’s results.

The solution is to make every execution use its own unique identifiers when interacting with anything outside n8n (files, storage paths, database keys, cache entries, etc.), At the same time, avoid global or static variables and keep data inside the execution context whenever possible, and if you expect many users running in parallel, add more compute or switch to a scaled/queue setup so executions don’t compete for the same limited resources.

3

This should not be happening, sharing your workflow would really help us narrow it down for you.

4

Hello Tamy. Thanks so much for the reply.

Will it be easy for you to advise me how to add more compute? or switching to a scaled/queue steup?

5

Additional info:
We are using Form Submission as trigger for execution.