Why Does the Freshservice Connector in Pipedream Only Fail with the Correct Subdomain?

This topic was automatically generated from Slack. You can find the original thread here.

Pipedream Freshservice connector — connection fails only with the correct subdomain

We’re hitting an issue trying to connect a customer’s Freshservice account through Pipedream’s Connect flow.

Symptom:
• Entering the customer’s correct Freshservice subdomain → Pipedream returns “Connection failed.”
• Entering any incorrect subdomain (typos, made-up strings, trailing punctuation) → Pipedream accepts the connection and saves it.
Only the correct subdomain fails. Every wrong one is accepted.

What we’ve already ruled out (Freshservice side):
Using the same API key the customer pasted into Pipedream, we ran direct HTTP calls from our laptop to four Freshservice read endpoints:
• GET /api/v2/agents/me
• GET /api/v2/tickets?per_page=1
• GET /api/v2/requesters?per_page=1
• GET /api/v2/departments?per_page=1
All returned HTTP 200 with valid data. Auth is standard Freshservice basic auth (API key as username, literal “X” as password). The key, subdomain, scope, and permissions are confirmed correct.

What would help:

  1. Can someone on Pipedream’s side dig into what’s actually happening when the correct subdomain is submitted? We’d like to know the underlying HTTP status / response body Pipedream is seeing from Freshservice, or a request ID we can correlate. We’re in contact with the Freshservice support team as well but can’t diagnose the issue without more detailed information.
  2. It would be much more debuggable for users if a failed validation surfaced the underlying error somewhere in the Pipedream dashboard. The current behavior - wrong subdomains pass validation, the correct one fails opaquely - is hard to diagnose without dropping down to direct API calls.
    Happy to share request/response details, the customer’s subdomain, or any other context over DM.

Bumping this

Missed this - will attempt to reproduce this and look into it as soon as possible.

thank you!

Can you confirm if this still occurs? I see Freshservice had an issue that was fixed recently - on June 3rd to be exact - and the behavior it would’ve exhibited matches exactly what you described