All-purpose, jobs and pipeline compute running in your own Azure subscription, which you set up and manage yourself. Its counterpart is serverless compute, which Azure Databricks runs for you.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Classic compute in context, with comparison tables and the common traps.
Terms in this definition
- Pipeline
Groups activities logically so that, together, they move and transform data, either one after another or side by side; found in both Azure Data Factory and Microsoft Fabric.
- Azure subscription
Container for Azure resources that also marks the edge of their billing, quotas and scale limits, governance, security and identity. Each one trusts a single Microsoft Entra tenant, and it isn't bound to any region.
- Set
Secret permission in Key Vault for writing secrets; some older material refers to it as Create.
- MANAGE
A Unity Catalog privilege allowing a principal to grant and revoke access on an object, hand over its ownership and drop it, all without being the owner. It gives no data access by itself and is not part of
ALL PRIVILEGES. - Serverless compute
Azure Machine Learning compute provided on demand whenever a job names no compute target. No cluster needs creating or managing, and jobs do not wait in a queue behind one another.
- Databricks
Analytics platform built on Apache Spark where data is processed in notebooks; Unity Catalog is the governance model it recommends.
- Agents (classic) API
First-generation Foundry Agent Service API, based on threads, messages and runs. It is deprecated, replaced by conversations and responses, and retires on 31 March 2027.
Related terms
- All-purpose compute
Interactive classic compute for notebook work, shareable between several people and restarted by hand. Its pricing differs from jobs compute, and production jobs are better run elsewhere.
- Auto termination
A classic compute option that shuts the compute down once it has been inactive for between 10 and 10,000 minutes; setting it to
0disables it. - Autoscaling
A classic compute option where you give a minimum and maximum worker count and Azure Databricks adjusts the number of workers within that range to match the load.
- Compute policy
Limits, set by an admin, on what can be chosen when people create classic compute; they help contain cost and require particular custom tags, libraries or init scripts.
- DBU
Databricks Unit: the billing measure for processing, charged by the second. The rate per unit varies by workload, pricing tier and VM size, and with classic compute you also pay Azure for the virtual machines themselves.
- Dedicated access mode
Assigns classic compute to one person or one group. Some capabilities work only in this mode, among them R, the RDD API, GPU instances and the ML flavour of Databricks Runtime.
- Driver node
In classic compute, the machine running the Spark driver: it keeps notebook state and collects results returned by
collect(), while workers host the executors. Unless you pick otherwise, its VM size matches the workers'. - Init scripts
Learn suggests steering clear of these where possible: shell scripts that execute on every classic compute node during boot, before Spark comes up. If you do need one, keep it cluster-scoped and store it in a volume; DBFS-hosted scripts are end-of-life.