Databricks Job ID's increasing in massive sequence gaps

[JoeWMP]

Has anyone seen something like this before? Today around midnight, our Job ID's started increasing in increments of quadrillions - was this a new change to how Job ID's are generated?

[JoeWMP]

If it helps, we've seen this in multiple Azure Databricks environments

[-werners-]

I guess that is possible, maybe databricks rolled out an update or changed the ID logic.

[JoeWMP]

Yeah, it can go up to int64 so they aren't invalid values, but it's very strange that it would increase in increments of 1 and now is jumping sequences in large increments. In an account that we had daily processing for a about a year, job id value only made it to around 4 million - now we're in the quadrillions or so

[Ravi · Accepted solution]

@Joe Swanson​  @Werner Stinckens​ Yes, we have started rolling out the job ID logic as mentioned here to all workspaces. https://docs.databricks.com/release-notes/product/2021/august.html#jobs-service-stability-and-scalab...

[JoeWMP]

Thank you Ravi! Glad that this confirms my understanding