My automation (trigger Telegram message, than AI, than Mongo) stuck and doesn’t work until I going to admin panel - stop & start again workflow.
I hadn’t any error messagers. Even in execution is - no log. Empty. No events.
After stop & start - I see - how to run workflow - I see result (All is ok) in executions log
Few things to help us figure more out for you:
Could you ctl a, ctl c the workflow, click the ‘</>’ button on the forums when drafting a reply and paste the workflow into the field that appears? Will help with context.
What/where is the docker compose hosted/located? What server specs does it have? Is it behind a domain with SSL? Just an IP?
(It sounds like a high load/inefficient workflow + low specs. But can’t say for sure.)
How? json is like 58000, but limit here is 32000…
Docker-compose - hosted on 1Gb Ram droplet on DigitalOcean. |
1 GB / 10 GB Disk / FRA1 - Ubuntu 24.10 x64
Have domain, ssl.
All cases works fine all time. But sometimes - one or two workflow - stuck and doesn’t work. After start/stop - handle all messages from bot chat - and normal processing it. Without any problems, or bugs.
Oh wow. Yeah.
That sounds like both cases.
An extremely large, potentially inefficient workflow. And massively under-spec for such a large usecase. I could recommend some workflow improvements potentially, maybe share as a google doc?
I would highly advise getting at least 4gb perhaps even 8gb RAM. 1gb RAM is insanely small. I have allocated over 42gb from my main server to n8n (rare case, handling millions of rows).
Thanks for your interest.
|My_Plan_Assist_copy.json - Google Drive
Added a temporary solution.
Created a flow that stops/starts other flows once an hour. If the flow itself doesn’t die, it will work )))
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.