I am running n8n in cluster mode, and I have an issue when starting n8n where telegram triggers would throw an error of too many requests when activating the workflows after n8n restart. I managed to go around the error by setting a low number of instances in the cluster (<3), however, I would love to be able to do 10 or more as I have enough resources (60 CPUs and 150GB RAM). Is there a way to rate limit workflow activation using the telegram trigger?
EDIT: I found another issue with this setup. The Schedule Trigger fires one time per instance. So one of the workflows was triggered 3 times. Is there a way to prevent this? I found this article: Configuring queue mode | n8n Docs but it looks quite complex to set up. Any help is appreciated.
Cluster mode is a native feature of pm2 (PM2 - Cluster Mode). I am guessing yes, this would be like having many main instances unless implemented with queue. I do have redis already running for another app, but I am no expert at this I would need to get my developer’s help. I was wondering is anyone ever tried this approach and how did it go for them.
Hi, as far as I know. Multi-main needs a special license and setup. If not, things get messed up.
For your case it might be interesting to look at queue mode and offloading to dedicated workers and webhook instances and leave the main only for UI purposes (most likely avoiding all these problems)