Workflow run failed: OnRamp's CRM connection has stopped working

Updated

Applies to: Workflows triggered by Salesforce or HubSpot records

A workflow run failed and no project was created. The run shows this message, with your CRM's name in it:

OnRamp's <CRM> connection has stopped working, so this Workflow could not run and no project was created.

Why it happens

The authorization behind your organization's CRM connection was revoked or disconnected. The connection doesn't fix itself, and retrying won't help until someone reconnects it.

While the connection is down, OnRamp records a failed run for the CRM change and doesn't retry it. No project is created for that change. Each workflow the change could have started gets its own failed run. That includes workflows whose trigger conditions couldn't be checked and workflows whose conditions had already matched.

The failed run has Error status. Its failing step is named Trigger conditions, and that step's result is Error.

If the record already has a settled run for a workflow, no new failed run is recorded for that workflow.

Fix it

We'll get the workflow running again in two steps.

  1. Reconnect your CRM to OnRamp. For the steps, see Connect Salesforce to OnRamp or Connect HubSpot to OnRamp.

  2. Update the record in your CRM to run the workflow again.

The failed run doesn't count as the record's run. After you reconnect, the next update to the record is checked as normal.

Who hears about failed runs

The workflow's alert recipients get an email that the run failed, with the error message. Each recipient gets at most one failure alert per workflow per clock hour. Later failures in that hour are counted into the next alert.

You choose recipients in the workflow's settings, under Alert recipients. Only active internal users can be recipients. If nobody is chosen, no failure email goes out. People who can publish workflows can edit Alert recipients.

Still stuck?

Contact support with the workflow's name, the time of the failed run, and a link to the CRM record that should have started it.

More in Troubleshooting