I know the recommended configuration for scaling out n8n is with Docker. I’m also aware that n8n has a configuration where one instance acts as a broker and others act as workers. But I wonder if I can accomplish the same thing with EC2 instances and a load balancer.
I have n8n configured to run on an EC2 instance, with MYSQL as the database running on a separate Amazon RDS cluster. Am I going to run into any problems if I simply snapshot the n8n instance, spin up another one, and put them both behind a load balancer? (I understand that I’ll have to keep all instances in sync running the same version of n8n.)
Am I going to run into any problems if I simply snapshot the n8n instance, spin up another one, and put them both behind a load balancer?
Hi @Lee_S, this isn’t a supported scenario I am afraid. When it comes to scaling n8n you’d typically have one main instance and multiple workers (and webhook processors) as described here: Configuring queue mode - n8n Documentation
So your webhook processors could sit behind a load balancers, but you’d need one additional main instance.
Thanks. Before I saw this reply I figured I’d just try it and see what happened. Well, what happened was scenarios that were running fun just started failing frequently but unpredictably. So word to the wise: don’t do that.
I’ll figure out the queue mode configuration in the next few weeks. Thanks!
