I have multiple doubts, I am practicing with n8n to make a RAG system, I have noticed that using the web or cloud version, telegram and all the nodes work well, except for the AI node, in my case I want to use OLLAMA which I have hosted on my computer, I put http://localhost: 11434 or http://127.0.0.1: 11434 and it gives ECCONREFUSED, I tried putting the ip of my computer with the respective port and the same error persists, so investigating the forums I found that it is more comfortable to use docker and install n8n there, I went and now I receive multiple errors both with telegram and with the same ollama error, so I decided to delete the container and try from npm on my windows 11, to my surprise in local npm it worked perfectly, put in the ollama node http://127.0.0.1: 11434 and connect does not reject the connection but the telegram error is: Problem running workflow
Bad request - please check your parameters
Show Details Telegram Trigger: Bad Request: bad webhook: An HTTPS URL must be provided for webhook
It is still in both docker and npm, I searched the forums, in one of the answers from 2021 they said that the windows version of npm was not supported, which I recently corrected and is having updates, anyway, my questions are:
Why does the Ollama AI node of the cloud and docker version of n8n reject my connection?
Why does this Telegram error occur in both docker and npm windows? (I have seen several solutions and it is still present today, I would like to know how I can configure the n8n webhook step by step or what is the case)
I am sorry for the inconvenience and touching the same topic from several years ago.
What is the error message (if any)?
Node Telegram error : Problem running workflow
Bad request - please check your parameters
Show Details
Telegram Trigger: Bad Request: bad webhook: An HTTPS URL must be provided for webhook
Ollama IA node in cloud and docker error :
Couldn’t connect with these
settings
Retrying
ECONNREFUSED
Please share your workflow
Share the output returned by the last node
There is no result since it cannot be executed by the telegram node, the expected result would be waiting for writing to the bot chat.
topic update, I read the thread you gave me, but it was chaotic, I tried a thousand ways to configure and place the free domain that it says in the thread, with nginx and win-acme because in my case it is local, and in no way does it take the domain, for the telegram node to work, just like in docker I tried and it doesn’t work, it is chaotic to try to make or connect a domain for the telegram node to work, finally I read a little more and I only needed to enter the router and open the respective ports to receive a connection with the domain and other connections, which I do not agree with, I do not feel safe since it is supposed to be local.
The issue of configuring webhooks and such was totally complicated and somewhat chaotic for me, is there any other solution? why does this error occur with the telegram node?
I understand now, but I thought it could be done another way since the environments I usually work in, few allow opening ports, for example Node-Red, I have worked on it for 3 years, and I had never had to do something like that, but it has me hooked n8n and I would like to understand more, do you know of any other solution?
I suppose you could use something like: Cloudflare Tunnel · Cloudflare Zero Trust docs
n8n might also have something built in with --tunnel but I’ve never used it. And this is something you’ll want to look at. It is instruction on how you would use the cloudflare system.
If this answers your question, please mark as a solution!
So the best thing to do for local n8n is to set it up in the cloud, meaning it’s more cost effective and there’s less room for error?
If this is the case, I wish there were other solutions for the local and these errors, meaning that the local already came implemented with the tunnel or something else, I don’t know if I make myself clear, thanks for the help!! I like this community.
If you don’t think you’ll be a good admin, then yeah go with cloud, but if you feel up to the challenge and want to self host with no open ports to your LAN, its doable but takes a little work.
Well after months, (I know it took some time, but I wanted to make sure of things), I did the n8n installation again with ngrok, so that the weebhooks and the telegram error would be solved, everything seemed correct, until I click on the telegram test flow or test trigger, I send a message to the telegram bot, and absolutely nothing happens, it stays waiting, I don’t know why but I’m sending it messages so that the respective data comes out within the trigger but it stays like that.
I’m updating to close this post. After attempting multiple Ngrok restarts and reading several documents, I’ve attached the following: First, run Ngrok with the assigned free tunnel. Then, run N8N in another terminal. Before running it, assign the URL Ngrok gave to the weebhook and then run N8N.
Example: set WEBHOOK_URL=https://xxxx-xxxx-xxxx.ngrok-free.app
N8N
This was the solution that worked for me. I’m sorry for the inconvenience, and I’m very grateful to those who responded and provided all the documentation. Thanks to you, I was able to find this solution. What a great community!