Analyze and optimize Google Cloud resources

This document describes how to use the conversational assistant in Gemini Cloud Assist to analyze and optimize Google Cloud resources.

In Gemini Cloud Assist, you can use the Cloud Assist pane to write natural language statements or questions (called prompts) to get in-depth explanations, suggested actions, or guided workflows. These resources help you complete tasks quickly and efficiently without needing to be a cloud expert.

Before you begin

Open the Cloud Assist pane

  1. Go to any page in the Google Cloud console, such as the Gemini for Google Cloud page.

    Go to Gemini for Google Cloud

  2. In the Google Cloud console toolbar, click spark Open or close Gemini AI chat.

    The Cloud Assist pane opens and displays a few example prompts that you can click. The prompts vary depending on your current Google Cloud console page.

  3. Optional: Write natural language prompts for assistance.

Prompt Gemini Cloud Assist for assistance

This section includes a few example scenarios and related prompts that you can enter in the Cloud Assist pane.

Generate SQL queries

The following example shows how to generate SQL queries from the Cloud Assist pane while in BigQuery pages of the Google Cloud console:

  1. In the Cloud Assist pane, send the following prompt:

    Generate a SQL query to show me the duration and subscriber type for the ten
longest trips.

    In this example, Gemini Cloud Assist provides a specific SQL query that you can review and copy into the BigQuery query editor.

  2. Optional: Send a prompt to generate Python code:

    Generate Python code to query the bigquery-public-data.ml_datasets.penguins
table using bigquery magics.

Understand deployed assets and their recent changes

Your project must have an organization to understand deployed assets and their recent changes. Projects that don't have an organization are not supported.

The following example uses the Cloud Asset Inventory.

  1. In the Cloud Assist pane, send the following prompt: What are the names of my instances and their private interface names?

    Gemini Cloud Assist lists instances, their names, and private interface names.

  2. Optional: Send any of the following prompts for similar scenarios:

    • List the name and network name for the subnetworks.
    • List my artifact repo.
    • How many BigQuery datasets have a label whose value is 'foo'?
    • List all config changes in the last 24 hours.
    • List all users granted roles/bigquery.admin and include the attached resources.
    • List a BigQuery dataset organized by name and location.
    • Describe the configuration of "pubsub/topic/…"

Analyze database query and system performance

The following example shows how to get the query performance data from the Cloud Assist pane while on the following tabs of a specific database instance in the Cloud SQL page of the Google Cloud console:

  • Overview
  • System Insights
  • Query Insights
  • Query Details

  1. In the Cloud Assist pane, send the following prompt:

    What's the current CPU utilization of this database?

    The response includes a list of all the compliance issues for the database fleet with a table that breaks down each category and the resource count.

  2. Optional: Send any of the following prompts for similar scenarios:

    • Summarize database query trends by users.
    • Which of my queries have high latency for this database instance?
    • What is the slowest query for this database in this instance?
    • What are the wait times for the query id = in this database?
    • How many error log entries are there for this database instance?

Understand database fleet health and compliance

The following example shows how to get the database fleet health and compliance data from the Cloud Assist pane:

  1. In the Cloud Assist pane, send the following prompt:

    How many databases do I have in the US?

    Gemini Cloud Assist provides a list of all of your databases in the US.

  2. Optional: Send any of the following prompts for similar scenarios:

    • Where are my databases located?
    • What issues are affecting my databases?
    • Show me all databases running on Postgres.
    • What databases have NIST issues in Europe?

The following example uses Cloud Monitoring.

  1. In the Cloud Assist pane, send the following prompt: Which of my VMs have the highest utilization?

    Gemini Cloud Assist provides a quick analysis of the performance trends of resources such as traffic, throughput, and executions.

  2. Optional: Send any of the following prompts for similar scenarios:

    • What are the top 5 dataflow jobs with the highest estimated backlog processing time?
    • Total VM network bytes sent per second for each instance?
    • What's the QPS on my SQL databases?

Find recommendations and optimization opportunities

The following example uses Active Assist.

  1. In the Cloud Assist pane, send the following prompt: Give me recommendations to reduce my resource costs.

    Gemini Cloud Assist provides specific recommendations for estimated savings and cluster reliability.

  2. Optional: Send any of the following prompts for similar scenarios:

    • What are all the quotas for my project that are almost hitting the limit?
    • What BigQuery datasets should I partition or cluster?
    • Which of my VMs are idle?

What's next