Compute Engine rate quotas and system limits

This document lists the Compute Engine rate quotas and system limits that specify the number of requests you can make to Compute Engine API methods.

Quotas restrict how much of a Google Cloud resource your project can use. If you want to change your quota value to accommodate more resource use, you can request a quota adjustment. System limits are fixed constraints which cannot be increased or decreased. For more information, see the Cloud Quotas overview.

Rate quotas and system limits

Rate quotas (also known as API quotas) define the number of requests that can be made to the Compute Engine API. These quotas apply on a per-project basis. Each quota applies to a group of one or more Compute Engine API methods. When you use the gcloud CLI or the Google Cloud console, you are also making requests to the API and these requests count towards your rate quota. If you use service accounts to access the API, that also counts towards your rate quota.

Google enforces rate quotas at a per-minute interval for each group. If your project reaches the maximum number of API requests within 60 seconds, you must wait for that rate quota to refill before making more requests in that group. For example, if a project reaches a quota limit at 10:00:15, the quota refills at the start of the next synchronized interval, such as 10:01:00, rather than refilling immediately. If your project exceeds a rate quota, you receive a 403 error with the reason rateLimitExceeded. To resolve this error, wait a minute then try your request again—the quota should be refilled at the start of the next interval.

The number of requests that you can make to the Compute Engine API is defined by API quotas as described in the following tables. Each group is counted separately, so you can achieve the maximum quota in each group simultaneously.

Rate quotas and system limits for global methods

The following sections list the quotas that define the number of requests you can make to the global Compute Engine API methods.

Simplified API quotas and system limits for global methods

To improve the discoverability and manageability of Compute Engine API quota, Google Cloud has reduced the number of quota metrics by consolidating metrics into fewer quotas. These simplified quota metrics also offer higher limits for each quota.

To view a complete list of global methods and the quota metrics that track the cost of each method, see Compute Engine API quota metrics reference.

The following table lists the simplified API quotas and metrics for Compute Engine API:

Quota and description	Metric name
Read requests per minute (`GlobalReadsPerMinutePerProject`) Quota for global `get` and `list` methods.	`compute.googleapis.com/global_reads`
Write requests per minute (`GlobalWritesPerMinutePerProject`) Quota for global write methods that are not included under other quota metrics.	`compute.googleapis.com/global_writes`
Cache invalidation requests per minute (`GlobalCacheInvalidationRequestsPerMinutePerProject`) Quota for global `UrlMapsService.InvalidateCache` methods.	`compute.googleapis.com/global_cache_invalidation_requests`

The following table lists the system limits and metrics for Compute Engine API:

System limit and description Metric name

List usable requests per minute (GlobalListUsablePerMinutePerProject)
System limit for global ListUsable methods. compute.googleapis.com/list_usable_requests

System limit and description	Metric name
List usable requests per minute (`GlobalListUsablePerMinutePerProject`) System limit for global `ListUsable` methods.	`compute.googleapis.com/list_usable_requests`
Filtered list cost overhead (`ListRequestsFilterCostOverheadPerMinutePerProject`) System limit for `.list` and `.aggregatedList` global methods with filters. Google Cloud charges quota usage for this metric in addition to the quotas against the `compute.googleapis.com/global_reads` metric. You incur quota charges if there are more than 10k resources filtered out of the list requests. Compute Engine API rejects the list requests if you exceed this quota limit.	`compute.googleapis.com/filtered_list_cost_overhead` Rate per project: 750k resources filtered out of the list requests per region per minute. As the quota is charged for every 10k resources, you see the limit as 75 (750k/10k) when you query for this quota limit on the Google Cloud console. Note: To avoid quota issues, optimize your list request filters or use list filters on the client side.
Requests per minute (`GlobalRequestsPerMinutePerProject`) System limit for all global read and write requests.	`compute.googleapis.com/global_requests`

Filtered list cost overhead (ListRequestsFilterCostOverheadPerMinutePerProject)

System limit for *.list and *.aggregatedList global methods with filters.

Google Cloud charges quota usage for this metric in addition to the quotas against the compute.googleapis.com/global_reads metric. You incur quota charges if there are more than 10k resources filtered out of the list requests. Compute Engine API rejects the list requests if you exceed this quota limit.

compute.googleapis.com/filtered_list_cost_overhead
Rate per project: 750k resources filtered out of the list requests per region per minute. As the quota is charged for every 10k resources, you see the limit as 75 (750k/10k) when you query for this quota limit on the Google Cloud console.

Requests per minute (GlobalRequestsPerMinutePerProject)

System limit for all global read and write requests.

compute.googleapis.com/global_requests

Legacy API quotas and system limits for global methods

The following table lists all Compute Engine API quotas for global methods.

Quota and description	Metric name
Queries per minute (`defaultPerMinutePerProject`) Quota for global list and mutation methods that are not included under other quota metrics. The following `get` methods also use this `default` metric: `networkFirewallPolicies.get` `projects.getXpnHost` `projects.getXpnResources`	`compute.googleapis.com/default`
Read requests (legacy) (`ReadRequestsPerMinutePerProject`) Quota for global `*.get` methods.	`compute.googleapis.com/read_requests`
List requests (`ListRequestsPerMinutePerProject`) Quota for global `*.list` methods.	`compute.googleapis.com/list_requests`
Operation read requests (`OperationReadRequestsPerMinutePerProject`) Quota for `globalOperations.get` method.	`compute.googleapis.com/operation_read_requests`
Global resource mutation requests (`GlobalResourceWriteRequestsPerMinutePerProject`) Quota for `images.delete`, `images.deprecate`, `images.insert`, `images.setLabels`, `snapshots.delete`, `snapshots.insert`, `snapshots.setLabels`, `machineImages.insert`, and `machineImages.delete` methods.	`compute.googleapis.com/global_resource_write_requests`
Heavy-weight mutation requests (`HeavyWeightWriteRequestsPerMinutePerProject`) Quota for `patch`, `delete`, and `insert` methods for the `interconnects` resources.	`compute.googleapis.com/heavy_weight_write_requests`
Heavy-weight read requests (`HeavyWeightReadRequestsPerMinutePerProject`) Quota for `*.aggregatedList` methods.	`compute.googleapis.com/heavy_weight_read_requests`

The following quotas apply to global APIs with per method quotas:

Quota and description	Metric name	Default limit
License insert requests Quotas for `licenses.insert` method.	`compute.googleapis.com/license_insert_requests`	Quota per project (`LicenseInsertRequestsPerMinutePerProject`): 2.5 requests/second (150 requests/minute) Quota per day per project (`LicenseInsertRequestsPerDayPerProject`): 30 requests/day
Project set common instance metadata requests Quota for `projects.setCommonInstanceMetadata` method.	`compute.googleapis.com/project_set_common_instance_metadata_requests`	Quota per project (`ProjectSetCommonInstanceMetadataRequestsPerMinutePerProject`): 36 requests/minute

Rate quotas and system limits for regional and zonal methods

The following sections list all quotas that apply to methods that use regional metrics.

Simplified API quotas and limits for regional and zonal methods

The following table lists the simplified API quota metrics for Compute Engine API regional and zonal methods: To view a complete list of regional and zonal methods, and the quota metrics that track the usage of each method, see Compute Engine API quota metrics reference.

Quota and description	Metric name
Read requests per minute per region (`ReadRequestsPerMinutePerProjectPerRegion`) Quota for regional and zonal `get` and `list` methods.	`compute.googleapis.com/reads_per_region`
Write requests per region (`WritesPerMinutePerProjectPerRegion`) Quota for regional and zonal write methods that are not included under other quota metrics.	`compute.googleapis.com/writes_per_region`

The following table lists the system limits for Compute Engine API regional and zonal methods:

System limit and description Metric name

List usable requests per region (ListUsablePerMinutePerProjectPerRegion)
System limit for regional and zonal ListUsable methods. compute.googleapis.com/list_usable_requests_per_region

System limit and description	Metric name
List usable requests per region (`ListUsablePerMinutePerProjectPerRegion`) System limit for regional and zonal `ListUsable` methods.	`compute.googleapis.com/list_usable_requests_per_region`
Filtered list cost overhead per region System limit for `.list` and `.aggregatedList` regional and zonal methods with filters. Google Cloud charges quota usage for this metric in addition to the quotas against the `compute.googleapis.com/reads_per_region` metric. You incur quota charges if there are more than 10k resources filtered out of the list requests. Compute Engine API rejects the list requests if you exceed this limit.	`compute.googleapis.com/filtered_list_cost_overhead_per_region` Rate per project (`ListRequestsFilterCostOverheadPerMinutePerProjectPerRegion`): 750k resources filtered out of the list requests per region per minute. As the quota is charged for every 10k resources, you see the limit as 75 (750k/10k) when you query for this quota limit on the Google Cloud console. Note: To avoid quota issues, optimize your list request filters or use list filters on the client side.
Requests per minute per region (

Filtered list cost overhead per region

System limit for *.list and *.aggregatedList regional and zonal methods with filters.

Google Cloud charges quota usage for this metric in addition to the quotas against the compute.googleapis.com/reads_per_region metric. You incur quota charges if there are more than 10k resources filtered out of the list requests. Compute Engine API rejects the list requests if you exceed this limit.

compute.googleapis.com/filtered_list_cost_overhead_per_region
Rate per project (ListRequestsFilterCostOverheadPerMinutePerProjectPerRegion): 750k resources filtered out of the list requests per region per minute. As the quota is charged for every 10k resources, you see the limit as 75 (750k/10k) when you query for this quota limit on the Google Cloud console.

Requests per minute per region (