[Feat]: Add get_job_info in BigQuery toolset

[afleisc]

Often times users want to know information about particular BigQuery jobs including answers to the questions:

• What is the state of my job?
• How long did my job take?
• How many slots did my job use?
• What were the referenced tables in my job?

etc

[afleisc]

The tool should be pretty straightforward - would be located in the existing metadata tool file: https://github.com/google/adk-python/blob/main/src/google/adk/tools/bigquery/metadata_tool.py

client.get_job returns Job ref: https://cloud.google.com/bigquery/docs/reference/rest/v2/Job

Glad to help with this

[aaslam17]

+1 needs this for many BQ agent usecases.

[shobsi]

@afleisc, @aaslam17 how does the user journey look like with this tool? Is the user going to input a job id? To note, currently the agent doesn't automatically know the job id for the jobs run by the execute-sql tool.

[afleisc]

@shobsi Correct, we have a use case where a user already has a job ID that they would like to analyze, so the user is responsible for passing the fully qualified job id (e.g. project-id:region.job_id) to ADK.

This is useful for a role like a DBA who is responsible for measuring the compute cost of particular queries that they've already identified

[itodotimothy6]

+1

The ability to analyze or retrieve information about a job is useful in many cases, especially for optimization use cases.

In addition to the above requested information in the description, get_job_info should also return the query that was run.

[bryceagenonordstrom]

+1

As part of our performance and cost program we would like this added as well. Getting the job information is vital for automated optimization efforts.

[dongyuj1]

dongyuj will take a look for this