Identify idle Cloud SQL instances
The Cloud SQL idle instancerecommender helpsyou detect instances that might be idle and provides youinsights and recommendations to help you reduce costs. This page describes howthis recommender works and how to use it.
How it works
The Cloud SQL idle instance recommender analyzes the usage metrics of primaryinstances that are older than 30 days. For each instance, the recommenderconsiders the values of certainmetricswithin an observation period spanning the last 30 days. The recommender does notanalyze read replicas.
If the activity level within the observation period is below a certainthreshold, the recommender estimates that the instance is idle. Recommendations aregenerated every 24 hours for shutting down such instances.
Warning: It's highly recommendedthat you evaluate the insights and in-context charts. Consult with theappropriate team to determine whether the flagged instance is inadvertently idleand can be deleted.Pricing
The Cloud SQL idle instance recommender is in theStandardRecommender pricing tier.
Before you begin
Before you can view recommendations and insights, do the following:
- To get the permissions to view and work with insights and recommendations,ensure that you have the requiredroles.
For more information about roles, seeunderstanding roles andgranting IAM permissions.Tasks Roles View recommendations One of these roles: recommender.cloudsqlViewerorcloudsql.viewer.Apply recommendations One of these roles: recommender.cloudsqlAdmin,cloudsql.editor, orcloudsql.admin. Enable the Recommender API.
List idle Cloud SQL instance recommendations
You can list idle instance recommendations byusing the Google Cloud console,gcloud, or the Recommender API.
Console
To list recommendations about idle instances by using theGoogle Cloud console, follow these steps:
Go to theCloud SQL Instances page.
ClickView all on the idle instance recommendations banner.
Alternatively, follow these steps:
Go to theRecommendation Hub.
For more information, seeGetting started with Recommendation Hub.
In theReduce Cloud SQL instance cost card, clickView all.
Click theShut down instance tab.
gcloud
To list recommendations about idle instances by usinggcloud, run thegcloud recommender recommendations listcommand as follows:
gcloud recommender recommendations list \--project=PROJECT_ID \--location=LOCATION \--recommender=google.cloudsql.instance.RECOMMENDER
Replace the following:
PROJECT_ID: Your project ID.LOCATION: A region, such asus-central1.RECOMMENDER: The ID of the recommender asIdleRecommender.
API
To list recommendations about idle instances by using theRecommendations API, call therecommendations.listmethod as follows.
GET https://recommender.googleapis.com/v1beta1/projects/PROJECT-ID/locations/LOCATION/recommenders/google.cloudsql.instance.RECOMMENDER/recommendations
Replace the following:
PROJECT_ID: Your project ID.LOCATION: A region, such asus-central1.RECOMMENDER: The ID of the recommender asIdleRecommender.
If the recommender detects idle instances, it lists them in atable. Each row shows the instance name, a brief recommendation, the location,and the last refresh date.
View insights and detailed recommendations
You can view insights and detailed recommendations about idle instances byusing the Google Cloud console,gcloud, or the Recommender API.
Console
To view insights about an idle instance by using theGoogle Cloud console, click the recommendation button in the list of instances.
gcloud
To view insights and detailed recommendations about idle instances by usinggcloud, run thegcloud recommender insights listcommand as follows:
gcloud recommender insights list \--project=PROJECT_ID \--location=LOCATION \--insight-type=google.cloudsql.instance.INSIGHT_TYPE
Replace the following:
PROJECT_ID: Your project ID.LOCATION: A region, such asus-central1.INSIGHT_TYPE: The ID of the insight type asActivityInsight.
API
To view insights and detailed recommendations about idle instances by using theRecommendations API, call theinsights.listmethod as follows:
GET https://recommender.googleapis.com/v1beta1/projects/PROJECT-ID/locations/LOCATION/insightTypes/google.cloudsql.instance.INSIGHT_TYPE/insights
Replace the following:
PROJECT_ID: Your project ID.LOCATION: A region, such asus-central1.INSIGHT_TYPE: The ID of the insight type asActivityInsight.
A panel appears showing insights about the idle instance. The followingcategories of information are provided along with some supporting charts.
| Fields | Descriptions and related metrics |
|---|---|
| Disk usage (bytes change) | Data utilization in bytes. Sampled every 60 seconds. Pertains to database/disk/bytes_used. |
| Max CPU utilization | The maximum CPU utilization as a percentage. Pertains to database/cpu/utilization. |
| Max memory utilization | The maximum memory utilization as a percentage. Pertains to database/memory/utilization. |
| Max disk read ops count | The maximum number of read operations. Pertains to database/disk/read_ops_count. |
| Max disk write ops count | The maximum number of write operations. Pertains to database/disk/write_ops_count. |
| Observation start and end dates | The start and end dates of the 30-day observation period. |
The following table lists the insight and recommendation that the Cloud SQLidle instance recommender might generate to help you reduce costs. The subtypesare visible in thegcloud and API results.
| Insight | Recommendation |
|---|---|
| The instance has shown low levels of activity during the observation period. Subtype: LOW_ACTIVITY | Investigate to check if the instance can be shut down to reduce costs. Subtype: SHUTDOWN_INSTANCE |
Apply recommendations
Evaluate the recommendations carefully and do any of the following:
To examine the recommendation, clickView instance.
A banner appears at the top of the page providing a recommendation.
You might find that the instance is intentionally idle and you don't need to shutit down. If it's inadvertently idle, you canstopanddelete the instance. Stopping aninstance suspends instance charges. The instance data is unaffected,and charges for storage and IP addresses continue to apply.
Exercise caution whenstoppingordeleting an instance. To avoidthe loss of data,create a backup first.
To dismiss the recommendation so that it's de-emphasized and appears dimmed,clickDismiss.
To close the panel without applying or dismissing the recommendation, clickCancel.
You can alsoexport these recommendations to BigQueryand examine them by usingLooker Studio orLooker.
What's next
- Monitor disk availability
- Reduce overprovisioned Cloud SQL instances
- Blog: Maximize your Cloud ROI
- Blog: Do more with less
- Google Cloud recommenders
Except as otherwise noted, the content of this page is licensed under theCreative Commons Attribution 4.0 License, and code samples are licensed under theApache 2.0 License. For details, see theGoogle Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.
Last updated 2025-07-14 UTC.