Scaling question

My project uses Webhook node as trigger, Postgres (Digital Ocean managed db) to store data, wait node for messages delay, and HTTP requests to send messages to Twilio or Telegram.

I’m running n8n with main option set, with Postgres, hosted at DO with Docker (1GB RAM, 1CPU, 25GB Disk).

Last 14 days, CPU usage is at 2% and memory around 60%.

This usage is for 1 bot. I plan to scale to 100 bots in the future.

My question is, for this type of application that is sensible to response time (Telegram / Twilio bots), what is the best scaling option?

Should I upgrade to a server with 16GB and 4CPUs using main option or, should I implement the redis with workers and/or webhook workers?

Thank you!

I would for sure use then the proper scaling setup with Redis.

Esp. because for a fast response you need to run n8n in ‘main’ mode and so only 1 CPU would be used and the other 3 would stay idle.

2 Likes