跳到正文
DEV Community· Desire Agency·· 7 小時前AI 評分60

n8n 初學者的三個內置節點工作流程及容易出錯之處

Three beginner n8n workflows with only built-in nodes, and the small things that broke first

AI 導讀

作者提供三個可直接匯入、只用 n8n 內置節點的初學者工作流程,涵蓋 RSS 摘要電郵、表單寫入 Google Sheet 和網站正常運作檢查。文章指出 RSS 空源仍應寄送提示電郵、表單欄位名稱須與節點中的表達式一致;網站檢查則在首次失敗後等待 60 秒再請求,只有第二次仍失敗才發出警報。

正文

I wanted three n8n workflows a beginner can import and run without installing a single community node. An RSS digest, a form that writes to a Google Sheet, and a website uptime check. All three are free on GitHub: https://github.com/MaximusSoroka/maximussoroka.github.io/tree/main/templates

None of them is clever. The useful part is the handful of details that decide whether they work the first time.

1. RSS to daily digest email

Schedule Trigger, RSS Read, Code, Send Email. Once a day it builds an HTML list of up to 20 items and mails it.

Two details:

  • The Code node escapes titles and links before building the HTML. Feeds contain & and < more often than you'd think.
  • An empty feed still sends a short "no items" email. A digest that silently stops arriving looks exactly like a digest that broke.

2. Form to Google Sheet plus notification

Form Trigger, Google Sheets, Send Email. Name, Email and Message go into a sheet tab called Submissions, then you get an email.

The trap: the form field labels have to match the expressions in the Sheet and email nodes. Rename "Message" to "Your message" in the form and the column stays empty, with no error. Change them in both places.

3. Website uptime check

Schedule Trigger, HTTP Request, If, Send Email. Every 5 minutes it requests a URL and alerts when the status isn't 200 or nothing comes back.

The detail that matters: by default the HTTP node throws on a 4xx/5xx or a timeout, and the workflow just stops. That's the exact case you want to catch. So the node uses full response, never error, and continue on error, and the If node decides.

What I got wrong: one failed request sends an alert. Two people on Reddit pointed out that a short network blip then looks like downtime. The fix they suggested is a Wait node and a second request before the email. That's in the template now: after a failed check it waits 60 seconds, requests the URL again, and only alerts if the second check fails too. Honest limit either way: this isn't a hosted monitor. If n8n is stopped, nothing checks anything.

Setup is the same for all three

Import workflow.json, swap the placeholder URL or sheet ID, attach your own credential (SMTP or Google), test against something you control, then activate. No credentials are in the files, only placeholder names.

If you run one of these and something breaks, I'd like to know where. Beginner setups fail in boring places, and boring places are what the README should cover.

This article was written with the help of AI and checked by me against the actual workflow files.

來源:DEV Community · dev.to