API Response Performance

[Mulen]

Hello.

I want to ask a theoretically question.

What is the minimum response time for a simple API respond workflow, n8n can achieve?

Let’s say you have a very permormant server, many cores, ram, fast internet and only one instance of n8n.

Then you have this simple Webhook Respond workflow.

What would be theoretically the fastest n8n can respond back? For example if i use insomnia client, the average response time is 70ms.

Is there a base “bottleneck” n8n has that what ever you do performance wise, n8n would not respond below 40ms for example?

What is n8n’s base performance for that simple workflow that can be achieved?

[Jon]

Hey @Mulen,

It all depends on what your worklfow is doing as that will be your limitation. If you were just returning a simple “hello” it would likely be a lot quicker, In my own testing on my personal n8n instance your example takes 70ms but a simple response of just { “status”: “ok” } takes 16ms.

[Mulen]

Hey @Jon thanks for the feedback.

Can you try the workflow i posted and tell me your average response time? I just want to create a baseline measurement to see what i have to tweak further.

[Jon]

It was around 70ms for my personal install going through an nginx reverse proxy and using a postgres database.

Another instance I have was ~40ms if that helps.