Exam DP-203 topic 4 question 17 discussion

It shall be A:Clusters, workspace logs does not have any cluster related resource change.

A is correct

ChatGPT : clusters Explanation: Clusters: In Azure Databricks, clusters are the key compute resource. Changes to clusters, such as creation, termination, and resizing, directly relate to compute changes. Logging actions related to clusters will help you track compute-related events. Workspace: Logging related to workspace typically involves monitoring user activities within the Databricks environment, such as notebook access, but it does not directly relate to compute changes. DBFS: The Databricks File System (DBFS) relates to storage, not compute resources. SSH: SSH logging is related to Secure Shell (SSH) connections, not specifically to Databricks compute resources. Jobs: While jobs involve executing tasks on clusters, they don't specifically log changes in the compute resources themselves. Thus, to monitor and log compute-related changes in Databricks, focusing on the clusters service is the appropriate choice.

You can't take logs from the workspace! Audit Logs are cluster-scoped: you need clusters.

Clusters are the most important service to log for changes in compute, as they represent the compute resources that are used to run workloads. By logging cluster-related actions, you can track when clusters are created, deleted, or modified. This information can be used to troubleshoot performance issues and ensure that compute resources are being used efficiently.

Answer : A.Clusters why not workspace ? Workspace is not a service that you should log to track changes in compute for the Databricks resource because it does not record events related to creating, editing, deleting, starting, or stopping clusters or jobs. Workspace events are related to actions performed on the workspace itself, such as creating, renaming, deleting, or importing notebooks, folders, libraries, or repos1. These events do not affect the compute resources used by the Databricks resource, but rather the workspace content and configuration. Therefore, workspace is not a relevant service for logging compute changes.

A is correct

https://learn.microsoft.com/en-us/azure/databricks/administration-guide/account-settings/audit-log-delivery MS document says cluster

Cluster is the only compute first class resource

A. Clusters: Databricks clusters are the primary compute resources in Azure Databricks. Monitoring and logging cluster-related actions will help you track changes in cluster creation, termination, resizing, and other cluster-related activities.

100% SURE A IS A CORRECT ANSWER.

definitely A

Workspace is correct. Detail is here: Set-AzDiagnosticSetting -ResourceId $databricks.ResourceId -WorkspaceId $logAnalytics.ResourceId -Enabled $true -name "<diagnostic setting name>" -Category <comma separated list> Link: https://docs.microsoft.com/en-us/azure/databricks/administration-guide/account-settings/azure-diagnostic-logs#configure-diagnostic-log-delivery

I thought compute is related to cluster.

Answer: A (clusters) Dispite using workspace to enable logging, from there you need to select clusters form the list if you want to satisfy the "changes in compute for the Databricks resource" question, hence the service you sould log is clusters. See link from Amsterliese. Beware of links to databricks.com vs links to microsoft because they are two slightly different products (i.e. Databricks (on AWS) vs Azure Databricks). For the other comment referencing dp200; the answer description only gives the defintions but no explanation.

From what I understand from MS documentation, it should be A - clusters https://docs.microsoft.com/en-us/azure/databricks/administration-guide/account-settings/azure-diagnostic-logs#configure-diagnostic-log-delivery The links in previous comments here which support answer B - workspace refer to AWS databricks. I tried to find a similar setup in the MS documentation, but couldn't find anything. Please tell me if my thinking is wrong. (Always happy to learn ;)

Agreed with clusters!