Workflow with 2 nested loops never completes resulting in Memory load on self-hosted server

Questions
1

Describe the problem/error/question

  • I run a workflow on scheduled trigger deployed in self-hosted server on Render
  • This workflow consists in:
    • getting data from an API by looping on page
    • for each page, loop over the items to store to Supabase
  • The data are correctly upserted to Supabase
  • But, the workflows never completes resulting in unclosed connections and memory load on the server

What is the error message (if any)?
image
image3024Ă—890 185 KB

Please share your workflow

Share the output returned by the last node

Last node is a Loop over items

Information on your n8n setup

  • n8n version: 1.105.4
  • Database (default: SQLite): PostgreSQL 16
  • n8n EXECUTIONS_PROCESS setting (default: own, main): main
  • Running n8n via (Docker, npm, n8n cloud, desktop app): Docker on render
  • Operating system: Linux srv-d1i24cumcj7s73eg1860-6fb6896788-qg9tx 6.8.0-1029-aws #31-Ubuntu SMP Wed Apr 23 18:42:41 UTC 2025 x86_64 Linux

More debug info —

instance information

Debug info

core

  • n8nVersion: 1.105.4
  • platform: docker (self-hosted)
  • nodeJsVersion: 22.17.0
  • database: postgres
  • executionMode: regular
  • concurrency: -1
  • license: enterprise (production)
  • consumerId: b042f870-50f9-4ca1-b952-1638ec1a0158

storage

  • success: all
  • error: all
  • progress: false
  • manual: true
  • binaryMode: memory

pruning

  • enabled: true
  • maxAge: 336 hours
  • maxCount: 10000 executions

client

  • userAgent: mozilla/5.0 (macintosh; intel mac os x 10_15_7) applewebkit/537.36 (khtml, like gecko) chrome/138.0.0.0 safari/537.36
  • isTouchDevice: false

Generated at: 2025-08-13T12:42:40.748Z

1 Like
2

Hi @Benoit_Cotte

The issue may be with the loop. From Executions, grab the previous run’s data, pin it to the starting node, and re-run the workflow. Watching the live execution will help you identify which loop is causing the problem.

4

Thanks a lot for you insight.

I watched the live execution, but given the large amount of items and could not reproduce the problem on tested items length.

I will retry with increasing successively both loops size.

5

UPDATE

I tried 2 things: 1) refactoring the workflow / 2) progressively increasing the number of items

About 1) Refactoring the workflow

  • First pass:
    • refactored the workflow so that “Loop over items” only have the API calls in the loop
  • Second pass:
    • Second “Loop over items” executed in a separated sub-workflow

Results:

  • Less memory usage
  • Higher impact on CPU spikes
  • no more server crashing (I run on Render, see above)

About 2) Increasing the number of items progressively

  • Workflow works well until processing 2600 items (in 5 minutes total)
  • After this, I have idle workflows never completing. It seems due to the second inner “loop over items” which deals with upserting to Supabase

From the latter bullet, I have a new hypothesis

  • this might be due to a connection issue with Supabase?
6

This is not solved yet. I wonder how can I bring more information to this problem.

7

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.