Loading

Considerations for using External System-Triggered Flows

Publiseringsdato: Jul 30, 2026
Beskrivelse

How External System-Triggered Flows Work

When an External System-Triggered Flow is activated:

  1. An initial watermark (typically a timestamp) is created to track the last successfully processed record.
  2. A background polling process is scheduled to periodically check the external system for new records using UC Connector (e.g., Microsoft Outlook connector).


During each polling cycle:

  • The poller queries records using a condition where last modified date > current watermark to fetch new records.
  • Connector may return zero or more new records (or events)
  • New records are processed one at a time.
  • Each record is published as a Platform Event (PE), which eventually triggers the Flow interview.
  • If PE published successfully: the watermark updates (to the last modified date of the processed record) and the poller waits for the next interval.

    Note: The watermark update behavior operates at the batch/page level. A failure on any single record within a page will prevent the watermark from advancing until that page is fully processed successfully.
  • If failed: the watermark is retained, and the poller retries up to 3 times. After 3 consecutive failures, the Flow is automatically deactivated.


Common Causes of Flow Deactivation

1. Unrecoverable systemic issues that require user intervention (e.g. connection expired and cannot be refreshed without user action)

2. Large Payload: if record size exceeds 128K character limit (it can happen more often with Outlook connector due to nature of the data returned single record can easily be longer than 128+K and cause flow de-activation)
If an incoming email's total payload exceeds this limit or there is any systemic issues, the Flow will fail and deactivate after 3 retry attempts.

Example Error message:

Retry Attempts Exhausted. Please fix the problem and reactivate the flow.
Error: externalservices.connector.poller.ExternalDataChgEventPublishException: 
Error [statusCode=STRING_TOO_LONG ...]
Løsning

1) If the issue is with Outlook connector where the payload size exceeds 128K character limit, 

  • Use Microsoft Power Automate to route incoming emails into two different folders based on the email body length.
    • If the email body exceeds the 125K character limit, move it to a "Rejected Emails" folder.
    • Otherwise, move it to an "Accepted Emails" folder.
  • On the Salesforce side, configure the Flow to poll only the following endpoint: {userId}/mailFolders/{AcceptedFolderId}
  • This ensures that Salesforce only processes email Folder with a body size below the 128K limit, preventing Flow failures caused by oversized emails.

2) Reactivate the Flow. If the error recurs consistently, contact Salesforce Support.

How to Reactivate a Deactivated Flow

  1. Go to Setup → Flows.
  2. Locate the deactivated Flow.
  3. Click Activate.
Flere ressurser

The 128K limit applies to the combined total payload, not any single field (including record metadata like field names).

Knowledge-artikkelnummer

005390149

 
Laster
Salesforce Help | Article