Power Platform Community Forum Thread Details

Environment

Dataverse (Dynamics 365) → ASL → ADLS Gen2 → Databricks → EDW

Sync interval: 30 minutes

Change Tracking: Enabled on all affected tables

Issue

We are using Azure Synapse Link to sync Dataverse tables to ADLS Gen2. We have a recurring issue where updates to existing records are not being synced while new record inserts sync correctly.

The issue occurs silently — no errors in EntitySyncFail folder, no alerts from ASL, and SinkModifiedOn continues showing recent timestamps giving a false impression that sync is healthy.

Business Impact:
A sales opportunity marked as Won in CRM continued showing as Open in our EDW due to the update never landing in ADLS, causing incorrect revenue reporting.

What We've Verified

Change Tracking enabled on all affected tables

ASL profile active and running

No errors thrown anywhere in the pipeline

Current Workaround

Unlink → Relink affected table → Full resync → Reprocess all layers. This fixes it temporarily but is not sustainable as we don't know when it will break again.

How can we proactively identify and resolve this synchronization issue before it impacts the reporting layer?

Thanks in advance.

Categories:

Microsoft Dataverse with Power Apps

Hi @CU06050840-0,

We’ve seen similar behavior before with Synapse Link, and the frustrating part is that the pipeline can look completely healthy while updates are silently being missed.

SinkModifiedOn is a bit misleading here because it only shows that ASL is still processing activity, not necessarily that actual row updates are making it into ADLS. So, the sync can appear “alive” even when update propagation has stalled.

One thing I’d strongly recommend is adding some reconciliation checks upstream instead of relying only on ASL health. Even something simple like comparing recent modifiedon values between Dataverse and ADLS can help catch the issue before it reaches reporting.

Unfortunately, your current workaround (unlink/relink/full sync) is still the standard recovery method most teams end up using when ASL gets into this state.

At this point, I’d recommend raising a Microsoft support ticket with a few concrete examples (record ID + modified timestamp in Dataverse vs missing update in ADLS). Silent update failures without anything in EntitySyncFail usually need backend investigation.

Please ✅ Does this answer your question if my post helped you solve your issue. This will help others find it more readily. It also closes the item.

If the content was useful in other ways, please consider answering Yes to Was this reply helpful? or give it a Like! 🩷