Queue Mode Sub-Workflow Has Added Latency

Questions
1

Describe the problem/error/question

I’m experiencing unexpected latency issues with sub-workflows when using n8n in Queue Mode. Here’s my current setup:

  • Dedicated Postgres instance
  • Redis
  • Webhook processor
  • Single worker instance
  • Frontend/test mode instance

All components are linked together within Queue Mode.

I’ve noticed that when I execute sub-workflows they appear to:

  1. Re-enter the queue
  2. Get executed the worker
  3. Then return results to the original workflow

This is causing significant latency:

  • Core logic execution within then sub-workflow (just one node): ~18ms
  • Complete round-trip when using sub-workflows: ~120ms

The latency seems to vary considerably between executions.

Are there any alternatives or workarounds to reduce this latency while still using sub-workflows?

Information on your n8n setup

  • n8n version: 1.90.2
  • Database (default: SQLite): PostgressSQL
  • n8n EXECUTIONS_PROCESS setting (default: own, main): queue
  • Running n8n via (Docker, npm, n8n cloud, desktop app): Docker
  • Operating system: Railway using all internal networking