09-12-2025 04:28 AM
Hi,
we have our Databricks Jobs deployed via DABs, and they have been running fine for a while now (approximately 1 month since we migrated from ADF). However, since yesterday, we are getting a weird issue while writing. See error below:
[STREAM_FAILED] Query [id = query_id, runId = run_id] terminated with exception: Do not know how to convert JArray(List(JString(some_key_string))) into class java.lang.String SQLSTATE: XXKSTFor context, we use AutoLoader File-Notification mode to fetch the data from ADLS, and store it into our bronze layer. Nothing changed - infrastructure wise - and I cannot find the reason why this is happening. It started in a couple of tasks - some working some not - and after a while, my whole bronze layer is not working anymore.
I would appreciate if you can nudge me in the right direction.
09-20-2025 10:20 AM
Hi @WiliamRosa,
not really. However, the issue resolved itself after 2 days. I guess it was because of an internal change (Microsoft’s or Databricks’ - not sure which). It appeared again one time last week, but I’m hoping it went away for good now.
09-12-2025 04:46 AM
Hi @BMex
There are more users reporting similar issues recently:
https://community.databricks.com/t5/data-engineering/intermittent-failures-with-auto-loader-on-azure...
I’ll wait for others in the community to help shed more light on this.
09-20-2025 09:33 AM
Hi @BMex,
The link I shared with a similar issue contains some solutions — did any of them work for you?
09-20-2025 10:20 AM
Hi @WiliamRosa,
not really. However, the issue resolved itself after 2 days. I guess it was because of an internal change (Microsoft’s or Databricks’ - not sure which). It appeared again one time last week, but I’m hoping it went away for good now.
Passionate about hosting events and connecting people? Help us grow a vibrant local community—sign up today to get started!
Sign Up Now