Task runner timeout on n8n Cloud - Code nodes failing for a week

Questions
1

Hi,
I’ve been getting “Task request timed out after 60 seconds” errors on Code nodes for over a week.
Instance: aiplug.app.n8n.cloud
Plan: Paid
Running via: n8n Cloud
The error says the task runner is “down, not ready, or at capacity.”
The n8n AI confirmed this is NOT a code issue - it’s infrastructure. My code is simple (extracting unique IDs from a Map).
What I’ve tried:

  • Reduced workflows from 90 to 34
  • Cleared old executions
  • Waited over a week
  • Tried different times of day
    I found threads about N8N_RUNNERS_TASK_REQUEST_TIMEOUT but I’m on Cloud and can’t set environment variables.
    Can someone help or restart my instance?
    Thanks!

Information on your n8n setup

  • n8n version:
  • Database: n8n Cloud (managed)
  • n8n EXECUTIONS_PROCESS setting: n8n Cloud (no access)
  • Running n8n via (Docker, npm, n8n cloud, desktop app): n8n cloud
  • Operating system: n8n Cloud (managed)
2 Likes
2

Hi @Flores_Shop_LLC

What is your n8n version?
Maybe updating to the latest version might fix the issue..

3

I’m currently on n8n version 1.121.3 (Cloud)

Also, I’ve noticed that some of my other nodes sometimes take a very long time to run or randomly break, which never happened before. The only consistent issue is with Code nodes, but overall the performance hasn’t been normal compared to before.

4

I had the exact same issue and the same n8n version, which is what made me update it to see if that’s what’s causing the problem. Now on version 1.122.5 (Cloud) it’s working again. That might just fix it for you as well.

5

This is quite annoying. My workflows keep failing randomly with an error

image
image1073×422 28.8 KB

I know I can configure the node to retry on fail, but this is not a sustainable, nor a long-term solution.

Do you know who can I contact from the n8n team to investigate this issue and provide a reliable solution?

6

any ideas to solution having the same problem :frowning:

7

I have the same issue since updating to 2.0.2 at a self-hosted instance. I thought it relates to the newly activated task runners but if there are other with same problems and 1.x versions it must be a different problem

8

I can confirm the same issue with self-hosted task runner as well,

code to reproduce:

image
image1029×486 42.3 KB

1 Like
9

+1 Same issue here. 2 workflows were running flawlessly yesterday. I upgraded to 2.14. from 1.x.x. Now both don’t work.

1 Like
10

+1 same issue here since I updated n8n to 2.14

11

+1 Same for me. Interestingly, manual starts of the workflows run without issues, it is just when the workflow is run normally (Webhook, Timer)

1 Like
12

For those having problems with code node when upgrading from v1 to v2 in self hosted versions, it is important to consider that there is a breaking change in v2: you should enable task runners and configure it correctly in docker compose file and in n8n-task-runners.json. If you plan to use Python, it is necessary to use external task runners and you will need to create a container for the task runners as well. It could be helpful to check for the container logs.

I have version 2.2.4 and finally managed to get the code nodes to work.

1 Like
13

I run into similar problems when running self-hosted n8n with docker images for n8n and n8n-task-runners. Could you give us more insights on what to focus or example compose/n8n-task-runner.json? For me it seems to be a problem to connect to redis, which could also a problem of misconfiguration in the json or the compose.yaml, so I would be very happy to compare with a working version.

14

+1

I’m having the same problem and changing environment variables that initially seemed to be related, such as N8N_RUNNERS_TASK_TIMEOUT, doesn’t solve the problem. I’m using version 2.X self-hosted with a task runner in a docker sidecar container.

15

+1, I’m seeing the same on self-hosted as well, but only in one specific workflow. Most of my other workflows run fine.

I already tried increasing N8N_RUNNERS_TASK_TIMEOUT, but that didn’t change anything.

What did help was reducing the amount of input data passed into the Code node by batching (using the Loop/Batching approach). In my case:

  • 1000 rows at once fails basically every time

  • ~500 rows sometimes fails

  • ~100 rows has been stable so far

I also checked the logs, but there’s nothing obvious or suspicious around the failures.

These are the task-runner-related environment variables I explicitly set:

N8N_RUNNERS_ENABLED=true
N8N_RUNNERS_MODE=external
N8N_RUNNERS_AUTH_TOKEN=${N8N_RUNNERS_AUTH_TOKEN}
N8N_RUNNERS_BROKER_LISTEN_ADDRESS=0.0.0.0
N8N_RUNNERS_TASK_REQUEST_TIMEOUT=120
N8N_RUNNERS_TASK_TIMEOUT=300
N8N_RUNNERS_MAX_CONCURRENCY=20
OFFLOAD_MANUAL_EXECUTIONS_TO_WORKERS=true
16

based on my case, I can confirm this is working without issues in V2.5 finally :slight_smile:

Edit:
note that my N8N_RUNNERS_TASK_TIMEOUT added in both n8n main & runner env vars,

after setting it to 300 and testing, it correctly stops when the code takes longer than expected,

I also believe this is a new error message, and it is well detailed:

image
image946×471 48 KB