IMAP trigger does not run automatically on n8n Cloud 2.39.6, but manual execution works

Questions
1

Describe the problem/error/question

I am using n8n Cloud. My published workflow does not execute automatically when new unread emails arrive in the configured IMAP inbox.
The workflow uses an Email Trigger (IMAP) connected to cp7145.webempresa.eu. When I execute the workflow manually, it retrieves the unread emails correctly. However, while the workflow is published, new emails remain unread in INBOX and no production execution is created.
Another published workflow in the same n8n Cloud instance uses a Schedule Trigger and executes successfully every 15 minutes. Therefore, production executions are working generally; the issue appears to be specific to the IMAP trigger.
I have already tried:

  • Setting “Force Reconnect Every Minutes” to 5
  • Unpublishing and publishing the workflow again
  • Sending a new email after publishing
  • Leaving the new email unread in INBOX
  • Waiting more than 10 minutes
  • Checking the production executions list
    The issue persists. Could this be a problem with the production IMAP listener or a known issue in n8n Cloud 2.39.6?
    Workflow name: Aurea — Email ingest (IMAP → ingest → CC+Vec)
    Workflow ID: C4UEy6Wjo3UMGluB

What is the error message (if any)?

There is no visible error message. The workflow remains published, but no automatic execution is created.
Manual execution successfully retrieves the same unread emails.

Please share your workflow

The relevant trigger configuration is:

{
  "parameters": {
    "options": {
      "customEmailConfig": "[\"UNSEEN\"]",
      "forceReconnect": 5
    }
  },
  "name": "Email Read IMAP2",
  "type": "n8n-nodes-base.emailReadImap",
  "typeVersion": 2
}

The trigger is connected directly to a Code node and then to an HTTP Request node.
I have not included credentials, API keys, or the complete production workflow because it contains private endpoint information. I can provide a sanitized workflow if needed.

Share the output returned by the last node

There is no output from a production execution because the automatic execution never starts.
During manual execution, the IMAP trigger returns the unread emails correctly and the downstream workflow executes.

Information on your n8n setup

  • n8n version: 2.39.6
  • Database (default: SQLite): Managed by n8n Cloud / unknown
  • n8n EXECUTIONS_PROCESS setting (default: own, main): Managed by n8n Cloud / unknown
  • Running n8n via (Docker, npm, n8n cloud, desktop app): n8n Cloud
  • Operating system: Managed by n8n Cloud
  • Instance ID: 2f994b2410e99effdd70a32084a3bf5abb772d78cb3c5a83343ef62a0b191b57
2

Hey @azelada7, while you wait for a response, here are some things that might help:

Suggested resources

Automatically matched to your question.

Docs:

Forum:

@Anshul_Namdev, @Gallo_AIA - you’ve helped with similar issues before, can you take a look?

Automatically suggested by n8n’s community bot. It’s a pilot - please share feedback here.

3

Manual execution runs a one-off search. The published trigger works differently: it keeps an IMAP connection open and waits for the server to push a new-mail notification (IMAP IDLE). Many shared cPanel hosts accept IDLE but never send that notification. When that happens the listener just waits, shows no error, and never starts an execution. Force Reconnect Every Minutes does not turn it into polling.

To confirm, open the same mailbox in Thunderbird and check whether new mail shows up right away or only when you refresh by hand. Also check that the credential uses port 993 with SSL/TLS on, because some hosts drop idle STARTTLS sessions on port 143.

If the server’s IDLE is broken, forward that inbox to Gmail or Outlook and use the Gmail Trigger or Microsoft Outlook Trigger instead. Both poll on a schedule.

Which port and SSL setting is the credential using?

4

Hi @azelada7, does manual execution also catch an email sent while it’s waiting, or only unread emails already in INBOX?

Temporarily unpublish the workflow, start with no unread emails in INBOX, then click Execute workflow and send a fresh email while it’s listening. Leave that email unread.

If it picks that up but the published workflow doesn’t, I’d focus on the production listener rather than assume Webempresa’s IDLE support is broken. Share that result and the test timestamp here so the n8n team can check it against your workflow ID.

5

Manual execution connects, searches for unread mail once and returns what’s there. Published, the trigger keeps a connection open and waits for the mail server to push a new-mail notice. If the server never sends one, or drops the idle connection without saying so, nothing fires. That matches what you’re seeing: no error, no execution, mail still unread.

cp7145 looks like a shared cPanel host, and that’s where I’d look first. Some of those have idle push switched off or throttled, and many cap simultaneous mail connections per user and address. n8n Cloud connects from shared addresses, so you can land on a limit you didn’t cause. Ask Webempresa support both questions.

To split host from n8n, duplicate the workflow, point the trigger at a Gmail or Outlook mailbox, publish it and send a test email. If that copy fires and yours doesn’t, it’s the host.

While you’re in there, delete the trigger and add a fresh email trigger node instead of editing the old one. You can clear the custom email config too, unread is what it searches for by default.

If the Gmail copy doesn’t fire either, say so here, that would point at the listener on Cloud rather than your mail server.