Predictive Optimization & Serverless Compute

[htd350]

Hello,

I have a hard time understanding how predictive optimization if serverless compute is not enabled. According to the documentation:

Predictive optimization identifies tables that would benefit from ANALYZE, OPTIMIZE, and VACUUM operations and queues them to run using serverless compute for jobs.

According to this discussion, having serverless compute enabled is not a prerequisite: Re: Databricks Predictive optimization - Databricks Community - 98408. If I don't have serverless workloads enabled how does predictive optimization optimize/analyze/vacuum the tables? We don't have it enabled because of our network situation.

Thanks in advance!

[Alberto_Umana]

Hi @htd350,

Predictive optimization in Databricks largely depends on the use of serverless compute to execute operations like ANALYZE, OPTIMIZE, and VACUUM, but not 100% sure if serverless is needed on all scenarios. I'll check internally and confirm on this, since documentation does not mention serverless as pre-requisite.