N8n Scaling without Redis

Every day I seem to find something new to automate with n8n : )
Makes me wonder what I should be looking out for when it comes to scaling our use of n8n.

running Redis seems a bit outside my skill set at this moment, but I can always upgrade my server as needed.
At this point, I have 2cpu’s and 4GB of ram and I don’t seem to be running into any issues.

What should I be looking out for when it comes to making sure I scale currency?
Can a go really far with 2cpu’s and 4GB of ram?
What are other ppl running?

I know there is no simple answer since each flow is different, I am just looking to get a better understanding as I use n8n more and more.

These questions are really hard to answer and will depend on your workflows and the data they process. In short: More nodes + more data would require more memory.

The best of course of action would be to run benchmarks. A very simple way if you want to process webhooks could just consist of running curl as often as needed and see if your instance handles it well enough (keeping an eye on the resource consumption - if anything maxes out you might need more). So something like this:

for i in {1..10}; do curl --request POST --url https://my.instance.url/webhook/my-path --header 'Content-Type: application/json' --data '{ "items": [ { "name": "Foo", "value": 10 }, { "name": "Bar", "value": 100 }, { "name": "Baz", "value": 1000 } ] }'; done

The above example should work on most Unix/Linux/MacOSX shells if curl is installed. Replace 10 with the amount of requests you want your instance to handle. With Windows/Microsoft Powershell you could put something similar together using Invoke-WebRequest.

As for what others are running, this is my n8n appliance:

That’s a Raspberry Pi 4 with 4 GB of RAM, I have yet to encounter actual issues coming down to the available resources here.

1 Like

The is everything in a nutshell.
Thank you so much for the sweet photos of your Raspberry Pi 4 n8n setup. I am thinking of doing a Raspberry Pi n8n sometime this year so I can play with everything locally.

1 Like

I also have n8n running locally on a Pi 4 with 4GB of RAM, I recently purchased some Nucs off ebay which have some older i5s in them with 8 and 16GB of RAM that I plan to move everything to.

I used the Pi for my personal automation tasks and as a dev environment for new workflows, It has held up nicely over the past few months and also runs nginx as a reverse proxy for other services I run from home.


Now that seems like a good idea when I really need to scale larger than my small DO server :slight_smile:

1 Like