Description of the problem:
When workflow is unavailable because of DNS resolution error, the error message shown is wrong
"Error: fetch failed"
instead of "Error during operation '${operation}' on workflow ${workflowId} with service URL ${serviceUrl}" with correct variable values

How reproducible:
100%

Steps to reproduce:

1. Setup dev env with cluster with orchestrator following https://docs.google.com/document/d/1EyQypjOFoiWrsiTAUcglwm_j_-pC8UkviavCOpCOxFA/edit?tab=t.0 
Don't add workflow URL to /etc/hosts

Actual results:
"Error: fetch failed" appears

Expected results:
"Error during operation '${operation}' on workflow ${workflowId} with service URL ${serviceUrl}" with correct variable values

The root cause: No try catch around call to nodejs  fetch
We are handling correctly the case where fetch returns a message with ok false, but not the case when fetch fails alltogether