Is it just me, or is the UI -painfully- slow?

I’m working on a project with a tight deadline, and I find myself waiting for the UI a LOT. I often get a “This page is not responding” popup in Chrome. For items with long inputs in the expressions field, it’s so bad that I’ve taken to copying the text into a text editor, working there, and then copying back.

In the past, I thought it was because I was accessing the UI through a load balancer and I figured the LB was screwing up something. But now I’m connected directly to the server, with no SSL, and it’s still really slow.

I tried changing the setting of N8N_PUSH_BACKEND, but that didn’t seem to help. Turning off saving manual executions helps with testing speed, but has a penalty that I can’t see what’s going wrong so it’s bad for diagnostics.

Amazon c7g.large instance running Amazon Linux 2023. Node 18. Database is Amazon Aurora Postgres RDS with scaling set to 0.5 to 4 ACUs. DB write instance and server are in the same zone (us-east-1b). Load on the server is running about 0.4-0.6, because I have some jobs already running.

Here’s an example of something I need to edit that’s driving me nuts:

I started with AWS for my n8n instance, but it was painfully slow and frustrating. So, I switched to Hetzner, which not only performs better but also costs less. Maybe you should think about giving it a shot too?

Usually the slowiness is because your browser consumes a big amount of memory. Even the EC2 t3.micro works well and fast with workflows.

The issue is also may be with the AWS region, if it uses something like US regions and you are located far away from it (like in Europe), there might be a small latency.

Plus, n8n may works slow if there is a lot of nodes in the workflow, like more than 50 nodes.