topic Re: How to get usage statistics from Databricks or SQL Databricks? in Warehousing & Analytics

How to get usage statistics from Databricks or SQL Databricks?

Mado — Thu, 20 Mar 2025 17:29:23 GMT

Hi,

I am looking for a way to get usage statistics from Databricks (Data Science & Engineering and SQL persona).

For example:

I created a table. I want to know how many times a specific user queried that table.
How many times a pipeline was triggered?
How long it took to run a DLT pipeline?

Is there any way to get usage statistics?

Re: How to get usage statistics from Databricks or SQL Databricks?

Anonymous — Tue, 14 Feb 2023 13:15:55 GMT

The Overwatch library may help. Unity Catalog also has some auditing to see who has accessed a table. For DLT, there are logs written to storage that you can read as a dataframe and parse for things like start and stop times and count how many starts to determine how often a pipeline is triggered.

Re: How to get usage statistics from Databricks or SQL Databricks?

youssefmrini — Thu, 23 Feb 2023 09:45:50 GMT

You can get those type of information by activating verbose audit logs. https://docs.databricks.com/administration-guide/account-settings/audit-logs.html

It contains a lot important metrics that you can leverage to build dashboards.

Re: How to get usage statistics from Databricks or SQL Databricks?

Afeez_Raheem — Thu, 23 Feb 2023 21:46:13 GMT

@Mohammad Saber , I hope that you are well. I have the same request from my manager. I was wondering if you found a way around this request already. We could work together, if you have not found the solution just yet

Please let's connect 🙂

Re: How to get usage statistics from Databricks or SQL Databricks?

Mado — Fri, 24 Feb 2023 01:14:19 GMT

@Owo Akilo

What I have found so far.

If Workspace is enabled for Unity Catalog:

I set up Azure Log Analytics following the documentation. Now, I get the logs in Azure Log Analytics Workspace.

There is a table named "DatabricksUnityCatalog" where I can find the action name "getTable". Here, I can see table names. But, for some reasons that I don't understand, Databricks doesn't save all table usage information in "DatabricksUnityCatalog".

I queried tables for a few days but I only see "getTable" action for one of the days.

If Workspace is not enabled for Unity Catalog:

What I found is:

1) Using cluster logs for queries run in a notebook (Ref).

After enabling cluster logs, I can see table names in a text file "log4j-active" but this file changes whenever I start the cluster. Seems that this file is archived as gz files but I cannot find table names in the archived files for most queries I had in the notebook.

2) Query history for queries run in the SQL persona (Ref).

We can see the query history in the Databricks user interface but I am looking for a way to send it to the Power BI, and extract table names.

I don't understand how API works and how to use API Specification file (Ref)

Re: How to get usage statistics from Databricks or SQL Databricks?

Afeez_Raheem — Fri, 24 Feb 2023 01:45:03 GMT

@Mohammad Saber Many thanks for sharing your learning thus far! I appreciate it

If you wouldn't mind, we could both connect on a Teams call to look into this more intimately. I am pretty new to learning the tool

Re: How to get usage statistics from Databricks or SQL Databricks?

Mado — Fri, 24 Feb 2023 02:08:06 GMT

@Owo Akilo

All right, you can see my email in my profile.

Re: How to get usage statistics from Databricks or SQL Databricks?

Afeez_Raheem — Fri, 24 Feb 2023 10:38:23 GMT

@Mohammad Saber thank you. I'd be sending you an email shortly

Re: How to get usage statistics from Databricks or SQL Databricks?

Afeez_Raheem — Fri, 24 Feb 2023 10:43:58 GMT

@Mohammad Saber , sorry it appears you don't have your email there. Would you help confirm, please?

Re: How to get usage statistics from Databricks or SQL Databricks?

twole — Thu, 16 May 2024 06:52:21 GMT

You can use System Tables, now available in Unity Catalog metastore, to create the views you described. https://docs.databricks.com/en/admin/system-tables/index.html