I am studying the course in n8n “Integrations: APIs & Connected Workflows” and I am working with the hands on exam but I am getting this error when running the step 1 test. I tested it in an online curl tool and it is working. See attached for your reference.
@rammendoza first off your webhook is actually fine, the 200 from your curl test proves the hostname resolves and its publicly reachable, so that error text is misleading, its just what the tester falls back to when its own call to your url doesnt come back clean. going by the testers own note, it calls your webhook using the same api key you log into the assessment platform with, so the fix is almost certainly making your Header Auth line up with that. open the “n8n Academy API Key” credential and check that the header Name and the Value are exactly what the platform expects, your curl passed because you sent your own matching header, but the tester sends the platforms key, and if your credential expects a different name or value the webhook rejects it and the tester reports it as not valid. one quick secondary thing, make sure the workflow is toggled Active, the production /webhook/ url only answers when its on, so if you tested then switched it off the tester would get nothing back and show the same message. is the key in that credential the same one you use to access the assessment platform?
Yes, it is the same. I just double-checked and ran the test again, and I got the same error. The workflow is published as well.
@rammendoza ok so not the auth then. quickest way to split this, when you hit test in the assessment tool, does a new execution show up in your n8n Executions tab? if nothing appears the testers request isnt reaching n8n at all (an edge thing, even though your curl works), and if an execution DOES appear then the request arrives fine and the tester is just unhappy with the response it gets back. also worth a quick confirm the webhook Method is POST and the url you gave the tester is the production one, not the /webhook-test/ one. which way does the executions tab go?
There is no execution showing up in the executions tab. I confirmed as well that the method is post and the url is the production (see screenshot above).
@rammendoza thats the key clue then, no execution means the testers request is dying before it ever runs your workflow, so its an edge/network thing and not your flow at all. on a published POST webhook that almost always comes down to one of two things, either the tester pings the url with a method other than POST first (n8n 404s those and logs no execution), or your hostinger instance is blocking the testers server (your own curl gets through but a datacenter IP doesnt). quickest way to tell them apart is to watch your n8n/hostinger logs while you run the test, if you see the request hit and 404 its the method, if you see nothing at all the tester cant reach the box. can you get at those logs?
@rammendoza the log explains it, your instance took a SIGTERM and restarted mid-session, so its cycling rather than staying up, and if the tester hits it during a down/restart window it just cant connect, which is exactly your “url invalid, no execution” (your own curl works because you happen to test while its up). on a small hostinger box thats almost always memory. theres also a separate flag, n8n is rejecting GET hits to the webhook (“did you mean POST”), so make sure the tester is POSTing, but id chase the restarts first.
Here is the n8n webhook tester
@rammendoza thats n8ns own academy tester but your log already pointed at your side, the restart plus the GET rejections. quick way to settle the GET part, temporarily add a second webhook node on the same path set to GET and re-run the tester, if it suddenly passes then the tester was probing with GET first. if it still fails its the restart, get the instance staying up reliably first (usually memory on that hostinger box).