Quotas and limits
- Why quotas are needed
- Default quotas and limits for Yandex.Cloud services
- Yandex Compute Cloud
- Yandex Object Storage
- Yandex Virtual Private Cloud
- Yandex Identity and Access Management
- Yandex Resource Manager
- Yandex Certificate Manager
- Yandex Key Management Service
- Yandex Network Load Balancer
- Yandex Container Registry
- Yandex Managed Service for Kubernetes
- Yandex Monitoring
- Yandex Managed Service for PostgreSQL
- Yandex Managed Service for ClickHouse
- Yandex Managed Service for MongoDB
- Yandex Managed Service for MySQL
- Yandex Managed Service for Redis
- Yandex Data Proc
- Yandex Message Queue
- Yandex Cloud Functions
- Yandex SpeechKit
- Yandex Translate
- Yandex Vision
- Yandex DataSphere
- Yandex DataLens
- Yandex IoT Core
- Yandex Cloud Interconnect
- Yandex Database
- Yandex Managed Service for SQL Server
- Yandex Managed Service for Apache Kafka®
- Yandex Managed Service for Elasticsearch
- Yandex API Gateway
- Cloud Logging
Yandex.Cloud services can be subject to quotas and limits:
- Quotas are organizational restrictions that can be changed by technical support on request.
- Limits are technical restrictions due to the Yandex.Cloud architecture specifics. The limits cannot be changed.
When designing your infrastructure in Yandex.Cloud, consider the limits as the maximum amount of resources that Yandex.Cloud can provide. Quotas are restrictions that can potentially be increased up to their limit.
Why quotas are needed
Quotas serve as a soft restriction for requesting resources and enable Yandex.Cloud to guarantee service stability, because new users can't take up too many resources for testing purposes. If you're willing to use more resources, you can increase them in the following ways:
- Submit an increase request.
- Contact the technical support and tell us which quotas you need to increase and by how much.
Default quotas and limits for Yandex.Cloud services
Quotas are listed with default values that match the quotas of the trial period.
Yandex Compute Cloud
Quotas
| Limits | Value |
|---|---|
| Number of virtual machines per cloud | 12 |
| Total number of vCPUs for all VMs per cloud | 32 |
| Total virtual memory for all VMs per cloud | 128 GB |
| Total number of disks per cloud | 32 |
| Total HDD storage capacity per cloud | 500 GB |
| Total SSD storage capacity per cloud | 200 GB |
| Total SSD storage capacity per cloud | 558 GB |
| Number of non-replicated disk placement groups per cloud | 5 |
| Total number of disk snapshots per cloud | 32 |
| Total storage capacity of all disk snapshots per cloud | 400 GB |
| Total number of file stores per cloud1 | 100 |
| Total file storage capacity per cloud1 | 0 B |
| Number of images per cloud | 8 |
| Number of images optimized for deployment per cloud1 | 0 |
| Number of instance groups per cloud | 10 |
| Total number of GPUs and vGPUs for all VMs per cloud1 | 0 |
| Number of concurrent operations per folder | 15 |
| Maximum number of placement groups per cloud | 2 |
| Number of dedicated hosts per dedicated host group1 | 0 |
| Number of dedicated host groups per cloud | 0 |
1 To increase quotas for file storages, VMs with GPUs, images optimized for deployment, or dedicated hosts, contact technical support.
VM limits
The limits per VM depend on its platform:
| Limits | Value |
|---|---|
| Maximum number of vCPUs per VM | Without vGPU: 32 With vGPU: 4 |
| Maximum virtual memory per VM | Without GPU and vGPU: 256 GB With GPU: 384 GB With vGPU: 12 GB |
| Maximum number of disks and file stores attached to a single VM2 | Less than 18 vCPUs: 8 From 18 vCPUs: 16 |
| Maximum number of GPUs connected to a single VM | 4 |
| Maximum number of vGPUs connected to a VM | 1 |
| Maximum number of security groups per interface | 5 |
| Maximum number of VM instances in a placement group | 5 |
| Limits | Value |
|---|---|
| Maximum number of vCPUs per VM | Without GPU: 80 With GPU: 64 |
| Maximum virtual memory per VM | Without GPU: 640 GB With GPU: 384 GB |
| Maximum number of disks and file stores attached to a single VM2 | Less than 20 vCPUs: 8 From 20 vCPUs: 16 |
| Maximum number of GPUs connected to a single VM | 8 |
| Maximum number of security groups per interface | 5 |
| Maximum number of VM instances in a placement group | 5 |
| Limits | Value |
|---|---|
| Maximum number of vCPUs per VM | 96 |
| Maximum virtual memory per VM | 640 GB |
| Maximum number of disks and file stores attached to a single VM2 | Less than 32 vCPUs: 8 From 32 vCPUs: 16 |
| Maximum number of security groups per interface | 5 |
| Maximum number of VM instances in a placement group | 5 |
2 Including the boot disk.
VM limits on operations with non-replicated disks
| Limits | Value |
|---|---|
| Total number of non-replicated SSDs per placement group | 5 |
| Maximum3 IOPS per vCPU | 10,000 |
| Maximum3 IOPS per VM | 100,000 |
| Maximum4 bandwidth per vCPU | 100 MB/s |
| Maximum4 bandwidth per VM | 1 GB/s |
Disk and file storage limits
| Limits | Value |
|---|---|
| Maximum disk size | 4 TB |
| Maximum disk snapshot size | 4 TB |
| Allocation unit size | 32 GB |
| Maximum3 IOPS for writes per disk | 40,000 |
| Maximum3 IOPS for writes per allocation unit | 1000 |
| Maximum4 bandwidth for writes per disk | 450 MB/s |
| Maximum4 bandwidth for writes per allocation unit | 15 MB/s |
| Maximum3 IOPS for reads per disk | 12,000 |
| Maximum3 IOPS for reads per allocation unit | 400 |
| Maximum4 bandwidth for reads per disk | 450 MB/s |
| Maximum4 bandwidth for reads per allocation unit | 15 MB/s |
| Limits | Value |
|---|---|
| Maximum disk snapshot size | 4 TB |
| Allocation unit size | 256 GB |
| Maximum3 IOPS for writes per disk | 11,000 |
| Maximum3 IOPS for writes per allocation unit | 300 |
| Maximum4 bandwidth for writes per disk | 240 MB/s |
| Maximum4 bandwidth for writes per allocation unit | 30 MB/s |
| Maximum3 IOPS for reads per disk | 300 |
| Maximum3 IOPS for reads per allocation unit | 100 |
| Maximum4 bandwidth for reads per disk | 240 MB/s |
| Maximum4 bandwidth for reads per allocation unit | 30 MB/s |
| Limits | Value |
|---|---|
| Minimum non-replicated disk size | 93 GB |
| Allocation unit size | 93 GB |
| Maximum3 IOPS for writes per disk | 75,000 (achievable for disks which use vhost, otherwise — at most 50,000) |
| Maximum3 IOPS for writes per allocation unit | 5600 |
| Maximum4 bandwidth for writes per disk | 1 GB/s |
| Maximum4 bandwidth for writes per allocation unit | 82 MB/s |
| Maximum3 IOPS for reads per disk | 75,000 (achievable for disks which use vhost, otherwise — at most 50,000) |
| Maximum3 IOPS for reads per allocation unit | 28,000 |
| Maximum4 bandwidth for reads per disk | 1 GB/s |
| Maximum4 bandwidth for reads per allocation unit | 110 MB/s |
| Limits | Value |
|---|---|
| Maximum storage size | 8 TB |
| Allocation unit size | 32 GB |
| Maximum number of files in storage | 1,000,000 |
| Maximum size of one file in storage | 8 TB |
| Limits | Value |
|---|---|
| Maximum storage size | 8 TB |
| Allocation unit size | 256 GB |
| Maximum number of files in storage | 1,000,000 |
| Maximum size of one file in storage | 8 TB |
Read and write operations utilize the same disk resource. The more read operations you do, the fewer write operations you can do, and vice versa. For more information, see Read and write operations.
3 To achieve maximum IOPS, we recommend performing read and write operations whose size is close to that of the disk block (4 KB by default).
4 To achieve the maximum possible bandwidth, we recommend performing 4 MB reads and writes.
Yandex Object Storage
Quotas
| Type of limit | Value |
|---|---|
| Storage volume in a cloud | 5 TB |
| Number of buckets per cloud | 25 |
Limits
| Type of limit | Value |
|---|---|
| Maximum object size | 5 TB |
| Total header size per request to the HTTP API | 8 KB |
| Size of user-defined metadata in an object | 2 KB |
| Maximum size of data to be uploaded per request | 5 GB |
| Minimum size of data parts for multipart uploading, except the last one | 5 MB |
| Maximum number of parts in multi-part uploading | 10,000 |
Yandex Virtual Private Cloud
Quotas
| Type of limit | Value |
|---|---|
| Number of cloud networks per cloud | 2 |
| Number of subnets per cloud | 6 |
| Number of all public IP addresses per cloud | 8 |
| Number of static public IP addresses per cloud | 2 |
| Number of routing tables per cloud | 8 |
| Number of static routes per cloud | 256 |
| Maximum number of security groups | 10 |
Limits
| Type of limit | Value |
|---|---|
| Minimum CIDR size for a subnet | /28 |
| Maximum CIDR size for a subnet | /16 |
| Maximum number of simultaneous TCP/UDP connections per VM | 50,000 |
| Supported network and transport layer protocols | IP, ICMP, TCP, UDP, GRE |
| Maximum number of security groups per interface | 5 |
| Maximum number of rules per security group | 50 |
| Maximum number of CIDRs per rule | 50 |
| Size of the DNS server IP address list | 100 characters |
Outgoing traffic filtering
Yandex.Cloud always blocks outgoing traffic to TCP port 25:
- To any servers on the internet other than Yandex.Mail servers.
- To Yandex Compute Cloud VMs when accessed via a public IP address.
Yandex.Cloud may open TCP port 25 by request via technical support if you comply with Acceptable Use Policy. Yandex.Cloud is entitled to block outgoing traffic on TCP port 25 if you violate the Use Policy.
Yandex Identity and Access Management
There are no quotas or limits for Yandex Identity and Access Management.
Yandex Resource Manager
There are no quotas or limits for Yandex Resource Manager.
Yandex Certificate Manager
Quotas
| Type of limit | Value |
|---|---|
| Number of certificates per cloud | 20 |
| Number of domains per cloud | 10 |
Limits for Let's Encrypt™ certificates
| Type of limit | Value |
|---|---|
| Number of domains per certificate | 100 |
| Number of certificates per domain | 50 per week |
| Number of certificates per set of domains | 5 per week |
See also
Yandex Key Management Service
Quotas
| Type of limit | Value |
|---|---|
| Number of keys per cloud | 100 |
Limits
| Type of limit | Value |
|---|---|
| Number of versions per key | 100 |
| Maximum amount of data to encrypt with the encrypt operation | 32 KB |
Yandex Network Load Balancer
Quotas
| Type of limit | Value |
|---|---|
| Number of load balancers per cloud | 2 |
| Number of target groups per cloud | 2 |
Limits
| Type of limit | Value |
|---|---|
| Number of resources per target group | 254 |
| Number of listening ports | 10 |
| Number of health checks per attached target group | 1 |
| Status check protocol | TCP, HTTP |
Other restrictions
A particular target group can only contain target resources from a single cloud network.
A target group can include resources that are connected to the same subnet within a single availability zone.
You can create a load balancer without a listener.
Health checks are transmitted from the IP address range 198.18.235.0/24.
When connecting resources to the load balancer, keep in mind the limit on the maximum number of simultaneous TCP/UDP connections per VM.
Yandex Container Registry
Quotas
| Type of limit | Value |
|---|---|
| Maximum number of registries per cloud | 10 |
| Number of simultaneous image scans running in a single cloud | 1 |
Limits
| Type of limit | Value |
|---|---|
| Maximum size of a single layer of a Docker image | 150 GB |
| Maximum number of Docker images in a single repository that can be checked by a trial run of the Docker image lifecycle policy | 50,000 |
Yandex Managed Service for Kubernetes
Quotas
| Type of limit | Value |
|---|---|
| Total number of vCPUs for all nodes per cloud | 192 |
| Total amount of RAM per cloud | 384 GB |
| Total disk storage capacity per cloud | 32 TB |
| Maximum number of Kubernetes clusters per cloud | 8 |
| Maximum number of node groups per cloud | 32 |
| Maximum number of nodes per cloud | 160 |
| Maximum number of node groups per Kubernetes cluster | 16 |
| Maximum number of nodes per Kubernetes cluster | 32 |
VM limits
| Type of limit | Value |
|---|---|
| Maximum number of volumes connected to a single node | 56 |
Yandex Monitoring
Quotas
| Type of limit | Value |
|---|---|
| Number of alerts per cloud | 10 |
General limits
| Type of limit | Value |
|---|---|
| Number of labels per metric (including mandatory ones) |
16 |
| Number of values for a single label | 1000 |
API limits
| Type of limit | Value |
|---|---|
| Maximum number of metrics returned in the data export API request | 500 |
| Maximum number of metrics sent in the data import API request | 10,000 |
| Minimum number of points returned in the data export API request | 10 |
| Maximum number of points returned in the data export API request | 10,000 |
| Minimum possible interval for exporting metrics | 1 second |
Alerting limits
| Type of limit | Value |
|---|---|
| Number of metrics per alert | 100 |
| Number of SMS per number per day | 50 |
Yandex Managed Service for PostgreSQL
Quotas
| Type of limit | Value |
|---|---|
| Number of clusters per cloud | 16 |
| Total number of processor cores for all DB hosts per cloud | 96 |
| Total virtual memory for all DB hosts per cloud | 640 GB |
| Total storage capacity for all clusters per cloud | 4096 GB |
Limits
| Type of limit | Minimum value | Maximum value |
|---|---|---|
| Host class | b1.nano (5% × 2 vCPU Intel Broadwell, 2 GB RAM) | m3-c80-m640 (80 vCPU Intel Ice Lake, 640 GB RAM) |
| Number of hosts in a cluster when using standard or fast network storage | 1 | 7 |
| Number of hosts in a cluster when using non-replicated network storage or fast local storage | 3 | 7 |
| Amount of data on the host when using fast network storage | 10 GB | 4096 GB |
| Amount of data on the host when using standard network storage | 10 GB | 2048 GB |
| Amount of data on the host when using non-replicated network storage | 93 GB | 8184 GB |
| Amount of data on the host when using local storage | 100 GB | 1500 GB |
Yandex Managed Service for ClickHouse
Quotas
| Type of limit | Value |
|---|---|
| Number of clusters per cloud | 16 |
| Total number of processor cores for all DB hosts per cloud | 96 |
| Total virtual memory for all DB hosts per cloud | 640 GB |
| Total storage capacity for all clusters per cloud | 4096 GB |
Limits
| Type of limit | Minimum value | Maximum value |
|---|---|---|
| Host class | b1.nano (5% × 2 vCPU Intel Broadwell, 2 GB RAM) | m3-c80-m640 (80 vCPU Intel Ice Lake, 640 GB RAM) |
| Number of ClickHouse hosts in a non-sharded cluster when using standard or fast network storage | 1 | 7 |
| Number of ClickHouse hosts in a non-sharded cluster when using non-replicated network storage | 3 | 7 |
| Number of ClickHouse hosts in a non-sharded cluster when using local storage | 2 | 7 |
| Number of shards in a sharded cluster | 1 | 50 |
| Number of hosts in a shard when using standard or fast network storage | 1 | 7 |
| Number of hosts in a shard when using non-replicated network storage | 3 | 7 |
| Number of hosts in a shard when using local storage | 2 | 7 |
| Total number of hosts per cluster | 1 | 353 (50 shards × 7 hosts + 3 ZooKeeper hosts) |
| Amount of data on the host when using standard or fast network storage | 10 GB | 4096 GB |
| Amount of data on the host when using non-replicated network storage | 93 GB | 8184 GB |
| Amount of data on the host when using local storage | 100 GB | 1500 GB |
Yandex Managed Service for MongoDB
Quotas
| Type of limit | Value |
|---|---|
| Number of clusters per cloud | 16 |
| Total number of processor cores for all DB hosts per cloud | 96 |
| Total virtual memory for all DB hosts per cloud | 640 GB |
| Total storage capacity for all clusters per cloud | 4096 GB |
Limits
| Type of limit | Minimum value | Maximum value |
|---|---|---|
| Host class | b1.nano (5% × 2 vCPU Intel Broadwell, 2 GB RAM) | m3-c80-m640 (80 vCPU Intel Ice Lake, 640 GB RAM) |
| Number of shards per cluster | 1 | 10 |
| Number of hosts in a shard when using network storage | 1 | 7 |
| Number of hosts in a shard when using local storage | 3 | 7 |
| Number of hosts per cluster | 1 host (clusters are unsharded at creation) | 70 (10 shards × 7 hosts) |
| Amount of data on the host when using network storage | 10 GB | 605 GB |
| Amount of data on the host when using local storage | 100 GB | 600 GB |
The maximum available number of simultaneous connections to a single Managed Service for MongoDB cluster host depends on the amount of RAM on the host:
| Amount of RAM | Maximum number of connections |
|---|---|
| 2 GB | 2048 |
| 4 GB | 4096 |
| 8 GB | 8192 |
| 16 GB or higher | 16,384 |
Yandex Managed Service for MySQL
Quotas
| Type of limit | Value |
|---|---|
| Number of clusters per cloud | 16 |
| Total number of processor cores for all DB hosts per cloud | 96 |
| Total virtual memory for all DB hosts per cloud | 640 GB |
| Total storage capacity for all clusters per cloud | 4096 GB |
Limits
| Type of limit | Minimum value | Maximum value |
|---|---|---|
| Host class | b1.nano (5% × 2 vCPU Intel Broadwell, 2 GB RAM) | m3-c80-m640 (80 vCPU Intel Cascade Lake, 640 GB RAM) |
| Number of hosts in a cluster when using standard or fast network storage | 1 | 7 |
| Number of hosts in a cluster when using non-replicated network storage or fast local storage | 3 | 7 |
| Amount of data on the host when using fast network storage | 10 GB | 4096 GB |
| Amount of data on the host when using standard network storage | 10 GB | 2048 GB |
| Amount of data on the host when using non-replicated network storage | 93 GB | 8184 GB |
| Amount of data on the host when using local storage | 100 GB | 1500 GB |
Yandex Managed Service for Redis
Quotas
| Type of limit | Value |
|---|---|
| Number of clusters per cloud | 16 |
| Total number of processor cores for all DB hosts per cloud | 64 |
| Total virtual memory for all DB hosts per cloud | 512 GB |
| Total storage capacity for all clusters per cloud | 4096 GB |
Limits
| Type of limit | Minimum value | Maximum value |
|---|---|---|
| Host class | b1.nano (burstable with 2 GB RAM) |
hm3-c32-m512 (32 vCPU Intel Ice Lake, 512 GB RAM) |
| Number of hosts in an unsharded cluster | 1 | 7 |
| Number of shards in a sharded cluster | 3 | 10 |
| Number of hosts in a shard when using network storage | 1 | 7 |
| Number of hosts in a shard when using local storage | 2 | 7 |
| Number of hosts in a sharded cluster | 3 (3 shards × 1 host) | 70 (10 shards × 7 hosts) |
| Amount of data on the host when using network storage | Twice the amount of RAM selected | 4096 GB |
| Amount of data on the host when using local storage | 100 GB | 1500 GB |
Yandex Data Proc
Quotas
VMs that make up Data Proc service clusters utilize your Yandex Compute Cloud cloud quotas.
Limits
| Type of limit | Value |
|---|---|
| Minimum host configuration | b1.small (20% × 2 vCPU Intel Broadwell, 4 GB RAM) |
| Maximum host configuration | s3-c96-m576 (96 vCPU Intel Ice Lake, 576 GB RAM) |
| Maximum number of hosts per subcluster | 32 |
| Maximum storage capacity for a cluster | 4096 GB |
Yandex Message Queue
Quotas
Messages
| Type of limit | Value |
|---|---|
Number of SendMessage and SendMessageBatch calls per queue |
300 calls per second for standard queues 30 calls per second for FIFO queues |
Number of ReceiveMessage calls per queue |
300 calls per second for standard queues 30 calls per second for FIFO queues |
Number of DeleteMessage and DeleteMessageBatch calls per queue |
300 calls per second for standard queues 30 calls per second for FIFO queues |
Number of ChangeMessageVisibility and ChangeMessageVisibilityBatch calls per queue |
300 calls per second for standard queues 30 calls per second for FIFO queues |
Number of CreateQueue calls per cloud |
2 calls per second |
Number of DeleteQueue calls per cloud |
5 calls per second |
| Number of other request calls per cloud | 100 calls per second |
| Number of queues per cloud | 10 |
Limits
Queues
| Type of limit | Value |
|---|---|
Minimum message enqueue delay (DelaySeconds parameter) |
0 seconds |
Maximum message enqueue delay (DelaySeconds parameter) |
900 seconds (15 minutes) |
| Number of messages being processed per standard queue | 120,000 |
| Number of messages being processed per FIFO queue | 20,000 |
| Queue name | Maximum of 80 characters, including numbers, small and capital Latin letters, hyphens, and underscores. The name of a FIFO queue must end with the .fifo suffix. |
Messages
| Type of limit | Value |
|---|---|
| Batch entry ID | Maximum of 80 characters, including numbers, small and capital Latin letters, hyphens, and underscores. |
| Maximum number of message attributes | 10 |
| Maximum number of entries per batch | 10 |
| Message content | XML, JSON, and unformatted text. The following Unicode characters are supported:
|
| Maximum period for retaining messages in a queue | 1209600 seconds (14 days) |
| Minimum period for retaining messages in a queue | 60 seconds (1 minute) |
Maximum message enqueue delay (DelaySeconds parameter) |
900 seconds (15 minutes) |
Minimum message enqueue delay (DelaySeconds parameter) |
0 seconds |
| Maximum message size | 262144 bytes (256 KB) |
| Minimum message size | 1 byte |
| Maximum message visibility timeout | 12 hours |
| Minimum message visibility timeout | 0 seconds |
Maximum clients waiting time to wait for the message from an empty queue (WaitTimeSeconds parameter) |
20 seconds |
Yandex Cloud Functions
Quotas
| Limits | Value |
|---|---|
| Number of functions per cloud | 10 |
| Number of function instances per availability zone | 10 |
| Number of concurrent function calls (for all functions in each availability zone) | 10 |
| Total RAM for all running functions per availability zone | 20 GB |
| Number of tags per function | 10 |
| Number of concurrent operations on a single function and its versions | 10 |
| Number of concurrent operations on all functions and their versions per folder | 15 |
| Number of concurrent operations on all functions and their versions per cloud | 20 |
A function instance processes one function call at a time. If there are more function calls than instances available, the call is queued. A queued call is treated as call-in-progress and consumes your quota.
Calls are distributed across availability zones randomly. Cloud Functions does not guarantee the even distribution of calls across zones. For example, all calls, regardless of their number, might end up in the same zone.
Limits
| Limits | Value |
|---|---|
| Maximum uploaded file size | |
| ZIP archive from the management console | 3.5 MB |
| ZIP archive from S3 bucket | 128 MB |
| Unzipped ZIP archive from S3 bucket | 680 MB |
| Other restrictions | |
| Maximum size of the JSON structure of a request and response | 3.5 MB |
| Maximum temporary file size | 512 MB |
| Maximum function execution time before timeout (including original initialization at first call) | 10 minutes |
| Maximum number of triggers per message queue | 1 |
Other restrictions
-
The trigger must be in the same cloud as the queue that messages are read from.
-
When installing dependencies:
- The total time for installing dependencies is limited to 5 minutes. After it expires, the build fails with an error.
- 1 GB of RAM is allocated to install dependencies, of which 700 MB is allocated for temporary file storage (tmpfs). Memory is consumed by active processes and files created and downloaded during the build.
- 50% of the CPU is available.
If the build takes longer than 5 minutes or requires more resources, gather the dependencies yourself and download the function as a ZIP archive.
Yandex SpeechKit
Quotas
| Type of limit | Value |
|---|---|
| Short audio recognition | |
| Requests per second | 20 |
| Streaming mode of short audio recognition | |
| Requests per second | 40 |
| Number of streams running simultaneously | 60 |
| Long audio recognition | |
| Recognition requests per hour | 500 |
| Operation status check requests per hour | 2500 |
| Billable hours of audio per day | 10,000 |
| Speech synthesis | |
| Requests per second | 40 |
Limits
| Type of limit | Value |
|---|---|
| Short audio recognition | |
| Maximum file size | 1 MB |
| Maximum duration of audio | 30 seconds |
| Maximum number of audio channels | 1 |
| Streaming mode of short audio recognition | |
| Maximum duration of transmitted audio for the entire session | 5 minutes |
| Maximum size of transmitted audio data | 10 MB |
| Maximum number of audio channels | 1 |
| Long audio recognition | |
| Maximum file size | 1 GB |
| Maximum duration of audio | 4 hours |
| Period for storing recognition results on the server | 3 days |
| Speech synthesis | |
| Maximum request size | 5000 characters |
Yandex Translate
Quotas
| Type of limit | Value |
|---|---|
| Calls of a single API method per second | 20 |
| Characters sent for translation per hour | 1 million |
| Characters sent for language detection per hour | 1 million |
Limits
For limitations on the field values in the request body, see the API reference.
The limits for using glossaries are as follows:
-
The number of pairs passed per glossary must be no more than 50.
-
The total number of characters in this structure must be no more than 20,000 in Unicode. 10,000 characters per source text and 10,000 characters per target text.
Warning
This only refers to texts without CGI parameters.
Yandex Vision
Quotas
| Type of limit | Value |
|---|---|
| Number of requests per second | 10 |
| Text recognition attempts per second | 5 |
| Face detection attempts per second | 5 |
| Image classification attempts per second | 5 |
Limits
| Type of limit | Value |
|---|---|
| Maximum file size | 1 MB |
| Maximum image size | 20 MP (length × width) |
| Maximum number of pages in a PDF file | 8 |
Yandex DataSphere
Quotas
| Type of limit | Value |
|---|---|
| Number of projects per cloud | 10 |
| Number of custom images per cloud | 3 |
| Number of datasets per project | 3 |
| Total dataset capacity per cloud | 500 GB |
| Project storage volume1 | 10 GB |
| Restrictions on the use of VM for individuals | |
| c1.32 | 0 |
| c1.80 | 0 |
| g1.1 2 | 0 |
| g1.2 | 0 |
| g1.4 | 0 |
| g2.1 | 0 |
| g2.8 | 0 |
1 for all projects created after October 15, 2021.
2 the quota increase is available when the balance of the billing account is more than $10 or by request to technical support.
Limits
Note
Technical restrictions on the maximum amount of data stored in the DataSphere project are valid until December 15, 2021 and apply only to projects created before October 15, 2021.
| Type of limit | Value |
|---|---|
| Maximum amount of data stored in DataSphere | 300 GB |
Yandex DataLens
General limits
| Type of limit | Value |
|---|---|
| Number of DataLens instances per cloud | 1 |
| Maximum CSV file size | 200 MB |
| Maximum number of rows of persisted data per dataset |
100,000,000 |
| Maximum number of tables per dataset | 32 |
Chart data limits
Each chart has a limit on the amount of data it can display.
Every time DataLens tries to visualize a chart, it makes a request to a dataset.
If the number of rows in the response exceeds the limit, DataLens doesn't display the chart.
You can change the number of data rows per response in request parameters, for example, by filtering.
| Chart type | Number of data rows from a dataset that the chart can display |
|---|---|
| Charts | |
| Line chart | 75,000 |
| Area chart Normalized area chart |
75,000 |
| Column chart Normalized column chart |
75,000 |
| Bar chart Normalized bar chart |
75,000 |
| Scatter chart | 75,000 |
| Pie chart | 1000 |
| Tree chart | 1000 |
| Tables | |
| Table | 100,000 |
| Pivot table | 40,000 |
| Geomaps | |
| Point map | 40,000 |
| Choropleth map | 40,000 |
| Heat map | 40,000 |
Yandex IoT Core
Quotas
| Type of limit | Value |
|---|---|
| Number of registries per cloud | 10 |
| Number of devices per cloud | 1000 |
Limits
| Type of limit | Value |
|---|---|
| Maximum size per message | 256 KB |
| Frequency of sending messages per device | No more than once per second |
| Maximum length of the subtopic name | 1024 symbols |
| Maximum number of aliases per device | 5 |
Note
Message delivery is not guaranteed if the limits are exceeded.
Other restrictions
- The
Retainflag is not supported. When using it, the MQTT broker saves the flagged message and sends it the next time you subscribe a device or registry to a topic. - Persistent Sessions, when the MQTT broker saves information about a client (device or registry) to reconnect faster, aren't supported.
Yandex Cloud Interconnect
Quotas
| Type of limit | Value |
|---|---|
| Number of private connections on a single trunk link | 2 |
Limits
| Type of limit | Value |
|---|---|
| The maximum number of prefixes that can be accepted by a cloud | 500 |
Yandex Database
Quotas
Dedicated instance mode
| Limits | Value |
|---|---|
| Total number of processor cores for all the database hosts in a cloud | 64 |
| Total virtual memory amount for all database hosts in a cloud | 256 GB |
| Maximum number of hosts per cloud | 8 |
| Maximum number of databases per cloud | 4 |
| Maximum number of storage groups per cloud | 8 |
Serverless mode
| Limits | Value |
|---|---|
| Maximum number of databases | 4 |
| Maximum number of serverless requests per second per cloud | 1000 RU |
| Maximum number of schema operations* per day per serverless database | 1000 |
| Maximum number of schema operations* per minute per serverless database | 30 |
*Transactions for which there is a quota: creating, updating, and deleting table schemas, getting a table description, getting a list of tables, and creating and deleting directories.
Things to note about applying quotas for the number of serverless requests per second
If the quota for the number of serverless requests is exceeded, a query is not accepted for execution and the Throughput limit exceeded error is returned. This error means that you can safely retry your query later. We recommend that you use the methods supplied as part of the YDB SDK when re-sending. The proposed methods implement retry algorithms that use different retry strategies depending on error type and code: zero delay, exponential backoff, fast or slow retry, and others.
The restriction is triggered with some delay, so the Throughput limit exceeded error is returned for a subsequent query rather than the specific query that resulted in exceeding the quota. After the restriction is triggered, requests will not be accepted for execution for a period approximately equal to the ratio of the requests over quota to the quota itself. If you use 1000 RU, for instance, in the execution of a single query while your quota is 100 RU, new requests will not be accepted for 10 seconds.
To reduce the risk of rejection under uneven load, YDB flexibly applies restrictions using a bandwidth reservation mechanism (burst capacity). As long as you use fewer RU processing requests than specified in the restriction, the unused bandwidth is reserved. During peak usage, more bandwidth than specified in the restriction may be temporarily available from the accumulated reserve.
YDB reserves about 5 minutes of bandwidth. The reserve enables you to run a single query with a cost of about 5 min × 60 s × quota RU/s without subsequent requests being rejected. The burst capacity availability policy may be changed.
The quota for the number of serverless requests is also a tool to protect from paying for resource overruns resulting from application faults or attacks on the service. The burst capacity mechanism enables you to set the quota to the lowest value at which your application works without getting any Throughput limit exceeded errors and to keep some margin against an unexpected increase in load.
Things to note about applying quotas for the number of serverless requests per second
If the quota for the number of serverless requests is exceeded, a query is not accepted for execution and the Throughput limit exceeded error is returned. This error means that you can safely retry your query later. We recommend that you use the methods supplied as part of the YDB SDK when re-sending. The proposed methods implement retry algorithms that use different retry strategies depending on error type and code: zero delay, exponential backoff, fast or slow retry, and others.
The restriction is triggered with some delay, so the Throughput limit exceeded error is returned for a subsequent query rather than the specific query that resulted in exceeding the quota. After the restriction is triggered, requests will not be accepted for execution for a period approximately equal to the ratio of the requests over quota to the quota itself. If you use 1000 RU, for instance, in the execution of a single query while your quota is 100 RU, new requests will not be accepted for 10 seconds.
To reduce the risk of rejection under uneven load, YDB flexibly applies restrictions using a bandwidth reservation mechanism (burst capacity). As long as you use fewer RU processing requests than specified in the restriction, the unused bandwidth is reserved. During peak usage, more bandwidth than specified in the restriction may be temporarily available from the accumulated reserve.
YDB reserves about 5 minutes of bandwidth. The reserve enables you to run a single query with a cost of about 5 min × 60 s × quota RU/s without subsequent requests being rejected. The burst capacity availability policy may be changed.
The quota for the number of serverless requests is also a tool to protect from paying for resource overruns resulting from application faults or attacks on the service. The burst capacity mechanism enables you to set the quota to the lowest value at which your application works without getting any Throughput limit exceeded errors and to keep some margin against an unexpected increase in load.
Yandex Managed Service for SQL Server
Quotas
| Type of limit | Value |
|---|---|
| Number of clusters per cloud | 16 |
| Total number of processor cores for all DB hosts per cloud | 96 |
| Total virtual memory for all DB hosts per cloud | 640 GB |
| Total storage capacity for all clusters per cloud | 4096 GB |
Limits
| Type of limit | Minimum value | Maximum value |
|---|---|---|
| Host class | s1.micro (2 vCPU Intel Broadwell, 8 GB RAM) | m3-c16-m128 (16 vCPU Intel Ice Lake, 128 GB RAM) for Standard Edition, m3-c80-m640 (80 vCPU Intel Ice Lake, 640 GB RAM) for Enterprise Edition |
| Number of vCPUs per cluster | 2 | 24 for Standard Edition and 448 for Enterprise Edition |
| Amount of RAM per cluster | 8 GB | 128 GB for Standard Edition, 4480 GB for Enterprise Edition |
| Number of hosts in a cluster when using standard or fast network storage | 1 | 1 for Standard Edition and 7 for Enterprise Edition |
| Number of hosts in a cluster when using non-replicated network storage or fast local storage | 3 for Enterprise Edition | 7 for Enterprise Edition |
| Amount of data on the host when using fast network storage | 10 GB | 4096 GB |
| Amount of data on the host when using standard network storage | 10 GB | 2048 GB |
| Amount of data on the host when using non-replicated network storage | 93 GB | 8184 GB |
| Amount of data on the host when using local storage | 100 GB | 1500 GB |
Yandex Managed Service for Apache Kafka®
Quotas
| Type of limit | Value |
|---|---|
| Number of clusters per cloud | 16 |
| Total number of processor cores for all broker hosts per cloud | 96 |
| Total virtual memory for all broker hosts per cloud | 640 GB |
| Total storage capacity for all clusters per cloud | 4096 GB |
Limits
| Type of limit | Minimum value | Maximum value |
|---|---|---|
| Host class | b2.medium (50% × 2 vCPU Intel Cascade Lake, 4 GB RAM) | m3-c80-m640 (80 vCPU Intel Ice Lake, 640 GB RAM) |
| Number of broker hosts in a cluster when using standard or fast network storage | 1 | 7 |
| Number of broker hosts in a cluster when using non-replicated network storage or fast local storage | 3 | 7 |
| Amount of data on the broker host when using fast network storage | 10 GB | 4096 GB |
| Amount of data on the broker host when using standard network storage | 10 GB | 2048 GB |
| Amount of data on the broker host when using non-replicated network storage | 93 GB | 8184 GB |
| Amount of data on the broker host when using local storage | 100 GB | 1500 GB |
Yandex Managed Service for Elasticsearch
Quotas
| Type of limit | Value |
|---|---|
| Number of clusters per cloud | 16 |
| Total number of processor cores for all the hosts per cloud | 96 |
| Total virtual memory for all hosts per cloud | 640 GB |
| Total storage capacity for all clusters per cloud | 4096 GB |
Limits
| Type of limit | Minimum value | Maximum value |
|---|---|---|
| Host class | s2.micro (2 vCPU Intel Cascade Lake, 8 GB RAM) | m3-c80-m640 (80 vCPU Intel Ice Lake, 640 GB RAM) |
| Number of hosts in a cluster when using standard or fast network storage | 1 | 7 |
| Number of hosts in a cluster when using non-replicated network storage or fast local storage | 3 | 7 |
| Amount of data on the host when using standard or fast network storage | 10 GB | 4096 GB |
| Amount of data on the host when using non-replicated network storage | 93 GB | 8184 GB |
| Amount of data on the host when using local storage | 100 GB | 1500 GB |
Yandex API Gateway
Quotas
There are no quotas for API Gateway.
Limits
| Type of limit | Value |
|---|---|
| Maximum size of a request to and a response from API Gateway | 2.5 MB |
| Maximum specification size | 3.5 MB |
| Maximum waiting time for a response from the service being integrated | 5 minutes |
Cloud Logging
Quotas
| Type of limit | Value |
|---|---|
| Number of custom log groups per cloud | 10 |
Limits
| Type of limit | Value |
|---|---|
| Adding records | |
| Maximum number of records per request | 100 |
| Maximum record size | 128 KB |
Maximum message size |
64 KB |
Maximum json-payload size |
64 KB |
| Maximum write speed | 10 MB/s or 1000 records per second |
| Reading records | |
| Maximum number of requests per second | 5 |
| Other restrictions | |
| Minimum record retention period | 1 hour |
| Maximum record retention period | 3 days |
| Maximum length of a filter | 1000 characters |