Forum Discussion
SQL Server CDC on VM DB Source Fails with "Property 'username' value '' is invalid" in Fabric Events
- 3 months ago
Hey v-dineshya ,
I wanted to share an update that I have successfully completed my Proof of Concept (POC) for streaming real-time data from SQL Server into Microsoft Fabric.
After evaluating a few approaches, I achieved the desired low-latency pipeline by implementing the following architecture:
SQL CDC → (Debezium + Kafka) → (Azure Event Hubs / Confluent.io) → Fabric Eventstream (via Event Hub or Confluent Connector)
Why this approach worked well:
Log-Based CDC: By using Debezium, the pipeline reads from the SQL transaction logs directly, ensuring zero impact on primary database query performance.
Scalability: Routing the data through Kafka/Event Hubs provides a highly reliable, buffered streaming layer.
Native Integration: Fabric’s native Eventstream connectors made it incredibly straightforward to ingest the stream directly into the lakehouse for real-time analytics.
Thanks to everyone for the guidance and support throughout this process! Hopefully, this architectural flow helps anyone else looking to implement a similar real-time CDC pattern.
Cheers!
Hi deborshi_nag,
Thank you for taking the time to respond to my query.
I reviewed the documentation you shared and noticed that the current connector support is limited to Azure VMs and similar cloud-hosted virtual machines.
As an alternative, I am currently exploring the following approach:
SQL Server CDC → Debezium → Kafka (Confluent) → Event Stream
I'm hoping this setup will work for my use case. Since I am learning and experimenting with the technology, I am also trying to leverage free trial options wherever possible.
Thanks again for your guidance. I will also try the public VM approach you suggested and will keep you posted on the results.
Your alternative architecture should work - you're running Debezium close to yuor laptop and using Kafka as a bridge to the cloud - let us know how it goes - good pattern when souce is not publicly accessible to cloud!