All Products
Search
Document Center

Elastic Compute Service:ECS Limits and Quotas

Last Updated:Apr 01, 2026

ECS enforces limits on resources, performance, and service usage. Some limits are fixed; others are adjustable by requesting a quota increase. Review these limits before deploying ECS to plan capacity, avoid hard constraints, and identify where to request increases.

How to read the tables

Each section contains a table with the following columns:

  • Limit: The specific constraint that applies, including fixed values or instructions for checking quota-based limits.

  • Adjustable: Whether the limit can be increased. If it can, the column contains a link to request an increase. If it cannot, the column shows "Not adjustable."

Some limits are quota-based — the actual value depends on your account and region. Use the quota ID to look up your current limit in the Quota Center.

A quota is the maximum amount of a cloud resource or the maximum number of operations that a single Alibaba Cloud account can use in a given region.

Instances

Restriction item

Limit

Adjustable

vCPU quotas

The maximum number of vCPUs per instance family per account per region, broken down by billing method (subscription, pay-as-you-go, or spot). See vCPU quotas below.

View or increase ECS instance type quotas

GPU and vGPU quotas

The maximum number of GPUs or vGPU instances per instance family per account per region, broken down by billing method. See GPU and vGPU quotas below.

View or increase ECS instance type quotas

Maximum subscription instances purchasable at one time (per account per region)

Quota ID: q_prepaid-instance-count-per-once-purchase. See View or increase ECS quotas.

Not adjustable

Convert pay-as-you-go to subscription

No limit on the number of instances. Discontinued instance types cannot be converted. See Convert a pay-as-you-go instance to a subscription instance.

Not adjustable

Convert subscription to pay-as-you-go

Each conversion generates a refund that counts against your monthly refund limit. If you exceed the limit, no further conversions are possible until the limit resets on the first day of the next month. See Convert a subscription instance to a pay-as-you-go instance.

Not adjustable

Secondary virtualization

Only ECS Bare Metal instances and Super Computing Clusters (SCC) support secondary virtualization. Other instance families do not support installing virtualization software or secondary virtualization.

Not adjustable

Sound card applications

Not supported.

Not adjustable

External hardware devices

Cannot directly attach external hardware devices such as hardware dongles, USB flash drives, external hard drives, or bank U-Keys. Software dongles or dynamic passwords can be used for secondary authentication instead.

Not adjustable

Multicast

Not supported. Use unicast point-to-point communication for one-to-many communication.

Not adjustable

ICP filing

To apply for an ICP filing, purchase a subscription instance with a duration of three months or more. Each instance supports ICP filings for a maximum of five websites or apps. See Check the server for ICP filing.

Not adjustable

Software licenses

Some software licenses are bound to instance hardware. Migrating an instance may change its hardware information and invalidate the license.

Not adjustable

vCPU quotas

Quota values are per region, not aggregated across all regions.

Quota ID

Quota description

Applicable instance families

Adjustable

q_ecs_restrict_prepay_c

Maximum vCPUs for subscription instances with purchase limits

xn4, n4, mn4, and e4 (previous-generation shared); Discontinued instance types

s6, shared standard instance family

s6, shared standard instance family

s6, shared standard instance family

View or increase ECS instance type quotas

q_ecs_local_storage_prepay_c

Maximum vCPUs for subscription instances with local storage (d and i series)

Big data instance families (d series); Instance families with local SSDs (i series)

View or increase ECS instance type quotas

q_ecs_share_prepay_c

Maximum vCPUs for subscription shared entry-level instances (e and t series)

e, economy instance family; t6, burstable instance family; t5, burstable instance family

View or increase ECS instance type quotas

q_ecs_ebm_prepay_c

Maximum vCPUs for subscription ECS Bare Metal instances (ebm, excluding GPU-accelerated types)

ECS Bare Metal instance types

View or increase ECS instance type quotas

q_ecs_hpc_prepay_c

Maximum vCPUs for subscription high-performance computing (HPC) instances (scc, excluding GPU-accelerated types)

SCC instance types

View or increase ECS instance type quotas

q_ecs_high_mem_prepay_c

Maximum vCPUs for subscription memory-optimized instances (re and se series)

re7p; re6p; re4; re4e; se1ne; se1

View or increase ECS instance type quotas

q_ecs_enterprise_prepay_c

Maximum vCPUs for subscription enterprise-level compute instances (g, c, r, u, hf, and sn series)

General-purpose (g series); Compute-optimized (c series); Enhanced (excluding re6p, re6, re4); Memory-optimized, r series (excluding se1ne, se1); Universal (U instances); High frequency (hf series)

View or increase ECS instance type quotas

q_ecs_restrict_postpay_c

Maximum vCPUs for pay-as-you-go instances with purchase limits

xn4, n4, mn4, and e4 (previous-generation shared); Discontinued instance types

View or increase ECS instance type quotas

q_ecs_local_storage_postpay_c

Maximum vCPUs for pay-as-you-go instances with local storage (d and i series)

Big data instance families (d series); Instance families with local SSDs (i series)

View or increase ECS instance type quotas

q_ecs_share_postpay_c

Maximum vCPUs for pay-as-you-go shared entry-level instances (e and t series)

e, economy instance family; t6, burstable instance family; t5, burstable instance family

View or increase ECS instance type quotas

q_ecs_ebm_postpay_c

Maximum vCPUs for pay-as-you-go ECS Bare Metal instances (ebm, excluding GPU-accelerated types)

ECS Bare Metal instance types

View or increase ECS instance type quotas

q_ecs_hpc_postpay_c

Maximum vCPUs for pay-as-you-go HPC instances (scc, excluding GPU-accelerated types)

SCC instance types

View or increase ECS instance type quotas

q_ecs_high_mem_postpay_c

Maximum vCPUs for pay-as-you-go memory-optimized instances (re and se series)

re7p; re6p; re4; re4e; se1ne; se1

View or increase ECS instance type quotas

q_ecs_enterprise_postpay_c

Maximum vCPUs for pay-as-you-go enterprise-level compute instances (g, c, r, u, hf, and sn series)

General-purpose (g series); Compute-optimized (c series); Enhanced (excluding re6p, re6, re4); Memory-optimized, r series (excluding se1ne, se1); Universal (U instances); High frequency (hf series)

View or increase ECS instance type quotas

q_ecs_restrict_spot_c

Maximum vCPUs for spot instances with purchase limits

xn4, n4, mn4, and e4 (previous-generation shared); Discontinued instance types

View or increase ECS instance type quotas

q_ecs_local_storage_spot_c

Maximum vCPUs for spot instances with local storage (d and i series)

Big data instance families (d series); Instance families with local SSDs (i series)

View or increase ECS instance type quotas

q_ecs_share_spot_c

Maximum vCPUs for spot shared entry-level instances (e and t series)

e, economy instance family; t6, burstable instance family; t5, burstable instance family

View or increase ECS instance type quotas

q_ecs_ebm_spot_c

Maximum vCPUs for spot ECS Bare Metal instances (ebm, excluding GPU-accelerated types)

ECS Bare Metal instance types

View or increase ECS instance type quotas

q_ecs_hpc_spot_c

Maximum vCPUs for spot HPC instances (scc, excluding GPU-accelerated types)

SCC instance types

View or increase ECS instance type quotas

q_ecs_high_mem_spot_c

Maximum vCPUs for spot memory-optimized instances (re and se series)

re7p; re6p; re4; re4e; se1ne; se1

View or increase ECS instance type quotas

q_ecs_enterprise_spot_c

Maximum vCPUs for spot enterprise-level compute instances (g, c, r, u, hf, and sn series)

General-purpose (g series); Compute-optimized (c series); Enhanced (excluding re6p, re6, re4); Memory-optimized, r series (excluding se1ne, se1); Universal (U instances); High frequency (hf series)

View or increase ECS instance type quotas

GPU and vGPU quotas

Quota values indicate the maximum number of GPUs or vGPU instances per account per region, not aggregated across all regions.

Quota ID

Quota description

Applicable instance families

Adjustable

q_ecs_ag_prepay_g

Maximum GPUs for subscription ARM-based GPU-accelerated instances

ebmgn6ia

View or increase ECS instance type quotas

q_ecs_vgpu_prepay_g

Maximum subscription vGPU-accelerated instances

sgn8ia; sgn7i-vws; vgn7i-vws; vgn6i-vws; vgn6i; vgn5i

View or increase ECS instance type quotas

q_ecs_gn7i_prepay_g

Maximum GPUs for subscription ebmgn7i/gn7s instances

gn7i; Elastic GPU Service (gn/vgn/sgn series); ebmgn7i

View or increase ECS instance type quotas

q_ecs_gn5_prepay_g

Maximum GPUs for subscription gn5 and gn5i instances

gn5; gn5i

View or increase ECS instance type quotas

q_ecs_gn6i_prepay_g

Maximum GPUs for subscription gn6i family instances

gn6i; ebmgn6i

View or increase ECS instance type quotas

q_ecs_gn6v_prepay_g

Maximum GPUs for subscription gn6v and ebmgn6v instances

gn6v; ebmgn6v

View or increase ECS instance type quotas

q_ecs_gn6e_prepay_g

Maximum GPUs for subscription gn6e and ebmgn6e instances

gn6e; ebmgn6e

View or increase ECS instance type quotas

q_ecs_gn8i_prepay_g

Maximum GPUs for subscription gn8is series instances

Elastic GPU Service (gn/vgn/sgn series); ebmgn8is

View or increase ECS instance type quotas

q_ecs_gn7v_prepay_g

Maximum GPUs for subscription gn7 and ebmgn7 instances

gn7; ebmgn7

View or increase ECS instance type quotas

q_ecs_gn7e_prepay_g

Maximum GPUs for subscription gn7e and ebmgn7e instances

gn7e; ebmgn7e

View or increase ECS instance type quotas

q_ecs_gn8v_prepay_g

Maximum GPUs for subscription gn8v and ebmgn8v instances

gn8v/gn8v-tee; ebmgn8v

View or increase ECS instance type quotas

q_ecs_ag_postpay_g

Maximum GPUs for pay-as-you-go ARM-based GPU-accelerated instances

ebmgn6ia

View or increase ECS instance type quotas

q_ecs_vgpu_postpay_g

Maximum pay-as-you-go vGPU-accelerated instances

sgn8ia; sgn7i-vws; vgn7i-vws; vgn6i-vws; vgn6i; vgn5i

View or increase ECS instance type quotas

q_ecs_gn7i_postpay_g

Maximum GPUs for pay-as-you-go ebmgn7i/gn7s instances

gn7i; Elastic GPU Service (gn, vgn, and sgn); ebmgn7i

View or increase ECS instance type quotas

q_ecs_gn5_postpay_g

Maximum GPUs for pay-as-you-go gn5 and gn5i instances

gn5; gn5i

View or increase ECS instance type quotas

q_ecs_gn6i_postpay_g

Maximum GPUs for pay-as-you-go gn6i and ebmgn6i instances

gn6i; ebmgn6i

View or increase ECS instance type quotas

q_ecs_gn6v_postpay_g

Maximum GPUs for pay-as-you-go gn6v and ebmgn6v instances

gn6v; ebmgn6v

View or increase ECS instance type quotas

q_ecs_gn6e_postpay_g

Maximum GPUs for pay-as-you-go gn6e and ebmgn6e instances

gn6e; ebmgn6e

View or increase ECS instance type quotas

q_ecs_gn8i_postpay_g

Maximum GPUs for pay-as-you-go gn8is series instances

Elastic GPU Service (gn/vgn/sgn series); ebmgn8is

View or increase ECS instance type quotas

q_ecs_gn7v_postpay_g

Maximum GPUs for pay-as-you-go gn7 and ebmgn7 instances

gn7; ebmgn7

View or increase ECS instance type quotas

q_ecs_gn7e_postpay_g

Maximum GPUs for pay-as-you-go gn7e and ebmgn7e instances

gn7e; ebmgn7e

View or increase ECS instance type quotas

q_ecs_gn8v_postpay_g

Maximum GPUs for pay-as-you-go gn8v and ebmgn8v instances

gn8v/gn8v-tee; ebmgn8v

View or increase ECS instance type quotas

q_ecs_ag_spot_g

Maximum GPUs for spot ARM-based GPU-accelerated instances

ebmgn6ia

View or increase ECS instance type quotas

q_ecs_vgpu_spot_g

Maximum spot vGPU-accelerated instances

sgn8ia; sgn7i-vws; vgn7i-vws; vgn6i-vws; vgn6i; vgn5i

View or increase ECS instance type quotas

q_ecs_gn7i_spot_g

Maximum GPUs for spot ebmgn7i/gn7s instances

gn7i; Elastic GPU Service (gn, vgn, and sgn); ebmgn7i

View or increase ECS instance type quotas

q_ecs_gn5_spot_g

Maximum GPUs for spot gn5 and gn5i instances

gn5; gn5i

View or increase ECS instance type quotas

q_ecs_gn6i_spot_g

Maximum GPUs for spot gn6i and ebmgn6i instances

gn6i; ebmgn6i

View or increase ECS instance type quotas

q_ecs_gn6v_spot_g

Maximum GPUs for spot gn6v and ebmgn6v instances

gn6v; ebmgn6v

View or increase ECS instance type quotas

q_ecs_gn6e_spot_g

Maximum GPUs for spot gn6e and ebmgn6e instances

gn6e; ebmgn6e

View or increase ECS instance type quotas

q_ecs_gn8i_spot_g

Maximum GPUs for spot gn8is series instances

Elastic GPU Service (gn/vgn/sgn series); ebmgn8is

View or increase ECS instance type quotas

q_ecs_gn7v_spot_g

Maximum GPUs for spot gn7 and ebmgn7 instances

gn7; ebmgn7

View or increase ECS instance type quotas

q_ecs_gn7e_spot_g

Maximum GPUs for spot gn7e and ebmgn7e instances

gn7e; ebmgn7e

View or increase ECS instance type quotas

q_ecs_gn8v_spot_g

Maximum GPUs for spot gn8v and ebmgn8v instances

gn8v/gn8v-tee; ebmgn8v

View or increase ECS instance type quotas

Image

Restriction item

Limit

Adjustable

Maximum custom images per account per region

Quota ID: q_user-image-count. See View or increase ECS quotas.

View or increase ECS quotas

Maximum accounts a single custom image can be shared with

Quota ID: q_user-per-image-shared-user-count. See View or increase ECS quotas.

View or increase ECS quotas

Image and instance type compatibility

Instance types with 4 GiB or more of memory cannot use 32-bit images.

Not adjustable

For more information, see Image overview.

Block storage

Block storage capacity uses binary units (base 1024). For example, 1 GiB = 1,024 MiB.

Pay-by-bandwidth: 200 Mbit/s

Pay-by-bandwidth:

  • Subscription instances: 200 Mbit/s

  • Pay-as-you-go instances: 100 Mbit/s

Pay-by-traffic:

  • Subscription instances: 200 Mbit/s

  • Pay-as-you-go instances: 100 Mbit/s

Pay-by-traffic: 100 Mbit/s

Restriction item

Limit

Adjustable

System disks per instance

1

Not adjustable

Data disks per instance

Varies by instance type. When creating an instance, you can specify up to 1 system disk and 64 data disks; the actual limit shown in the console applies. Attach additional data disks after creation if needed. Call DescribeInstanceTypes to query the limit for a specific type. See Instance families.

Not adjustable

Disk capacity quota per account per region and zone

Varies by disk type. See View or increase block storage quotas.

View or increase block storage quotas

Basic disk capacity

5 GiB to 2,000 GiB

Not adjustable

Standard SSD capacity

20 GiB to 32,768 GiB

Not adjustable

Ultra disk capacity

20 GiB to 32,768 GiB

Not adjustable

Enhanced SSD (ESSD) capacity

PL0: 1–65,536 GiB; PL1: 20–65,536 GiB; PL2: 461–65,536 GiB; PL3: 1,261–65,536 GiB

Not adjustable

ESSD AutoPL disk capacity

1 GiB to 65,536 GiB

Not adjustable

ESSD Entry disk capacity

10 GiB to 32,768 GiB

Not adjustable

Regional ESSD capacity

1 GiB to 65,536 GiB

Not adjustable

Local SSD capacity

Varies by instance type, from 5 GiB to 7,152 GiB per disk

Not adjustable

Total local SSD capacity per instance

Varies by instance type, up to 8 × 7,152 GiB

Not adjustable

Elastic ephemeral disk capacity

64 GiB to 8,192 GiB

Not adjustable

System disk capacity

Windows Server: 40–2,048 GiB; FreeBSD: 30–2,048 GiB; Other Linux: 20–2,048 GiB. When a basic disk is used as a system disk, the maximum capacity is 500 GiB.

Not adjustable

Attach new local disks to an instance that already has local disks

Not supported.

Not adjustable

Configuration changes for instances with local disks

Only bandwidth changes are supported.

Not adjustable

Linux system disk mount point

/dev/vda

Not adjustable

Linux data disk mount points

1–25 data disks: /dev/vd[b-z]. More than 25 data disks: /dev/vd[aa-zz] (for example, the 26th disk is /dev/vdaa, the 27th is /dev/vdab).

Not adjustable

For more information, see Block storage overview.

Snapshots

Restriction item

Limit

Adjustable

Manual snapshots per disk

2,000

Not adjustable

Automatic snapshots per disk

2,000

Not adjustable

Archived snapshots per disk

10,000

Not adjustable

Automatic snapshot policies per account per region

100

Not adjustable

Automatic snapshot policies per disk

10

Not adjustable

Concurrent snapshots per ESSD disk (ESSD, ESSD AutoPL, ESSD Entry, regional ESSD)

10

Not adjustable

Concurrent snapshots per legacy disk (standard SSD, ultra disk, basic disk)

1

Not adjustable

Concurrent archived snapshots per disk

10

Not adjustable

Snapshot support by disk type

Cannot create snapshots for local disks or elastic ephemeral disks. The snapshot-consistent group feature requires ESSD series disks (ESSD, ESSD AutoPL, or ESSD Entry) with the multi-attach feature disabled. See Multi-attach. The instant access (IA) feature requires ESSD series disks (ESSD, ESSD AutoPL, ESSD Entry, or regional ESSD). See Snapshot instant access.

Not adjustable

Download or export snapshots

Not supported directly. Create a custom image from the snapshot, then export the image to your on-premises device. See Create a custom image from a snapshot.

Not adjustable

Concurrent manual and automatic snapshots — ESSD series disks

Supported. If the concurrent snapshot limit is reached, subsequent snapshot tasks fail. See the concurrent snapshot limits above.

Not adjustable

Concurrent manual and automatic snapshots — legacy disks (standard SSD, ultra disk, basic disk)

Not supported. If a snapshot is in progress when an automatic snapshot is scheduled, the automatic snapshot is skipped and retried at the next scheduled time. If an automatic snapshot is in progress, wait for it to complete before creating a manual snapshot.

Not adjustable

For more information, see Snapshot overview.

Public bandwidth

Account-level total peak bandwidth

The following limits apply to the combined peak bandwidth of all pay-as-you-go and spot instances using the pay-by-bandwidth billing method under a single account in a single region.

Region

Total peak bandwidth

China (Beijing), China (Shanghai), China (Hangzhou), and China (Shenzhen)

50 Gbps

China (Hong Kong) and Singapore

20 Gbps

All other regions

10 Gbps

To increase the limit, use the quota ID q_internet-bandwidth-pay-by-bandwidth-of-postpaid-instance. See View or increase ECS quotas.

Starting November 27, 2020, account-level rate limiting policies also apply when creating or changing ECS instance configurations:

  • The combined actual peak bandwidth of all pay-by-traffic instances in a single region cannot exceed 5 Gbps per account.

  • The combined actual peak bandwidth of all pay-by-bandwidth instances in a single region cannot exceed 50 Gbps per account.

To request a higher peak bandwidth, submit a ticket.

Per-instance bandwidth limits

Important

For the pay-by-traffic billing method, peak bandwidth values are upper limits and are not guaranteed. During periods of resource contention, bandwidth may be reduced. Use the pay-by-bandwidth billing method if guaranteed bandwidth is required.

Restriction item

Limit

Adjustable

Peak inbound bandwidth

If the purchased outbound peak bandwidth is 10 Mbit/s or less, Alibaba Cloud allocates 10 Mbit/s inbound. If the purchased outbound peak bandwidth exceeds 10 Mbit/s, inbound bandwidth equals the purchased outbound bandwidth.

Not adjustable

Peak outbound bandwidth — pay-by-traffic

Subscription instances: 200 Mbit/s; pay-as-you-go instances: 100 Mbit/s. The total public bandwidth of an instance is also capped by the instance type's baseline network bandwidth (see the Baseline Network Bandwidth column in Instance families). Some instance types (ecs.t6-c4m1.large, ecs.t6-c2m1.large, ecs.t6-c1m1.large, ecs.t6-c1m4.large) are limited to 80 Mbit/s.

Not adjustable

Peak outbound bandwidth — pay-by-bandwidth

200 Mbit/s. The total public bandwidth of an instance is also capped by the instance type's baseline network bandwidth.

Not adjustable

Public IP address replacement

Replace the public IP address of a new instance within 6 hours of creation. Maximum 3 replacements per instance.

Not adjustable

For more information, see Public bandwidth.

Elastic Network Interfaces

Restriction item

Limit

Adjustable

Maximum secondary Elastic Network Interfaces (ENIs) per account per region

Quota ID: q_elastic-network-interfaces. See View or increase ECS quotas.

View or increase ECS quotas

VPC and zone requirements for ENI attachment

The instance and any attached ENI must be in the same Virtual Private Cloud (VPC) and the same zone. Multiple ENIs attached to an instance can belong to different vSwitches within the same VPC and zone. Attaching two or more ENIs from the same vSwitch to an instance may cause asymmetric routing issues. Assign secondary private IP addresses to an ENI instead. See Secondary private IP addresses.

Not adjustable

Maximum ENIs per instance

Determined by instance type. See the Elastic Network Interface column in Instance families.

Not adjustable

For more information, see ENI overview.

Prefix lists

Restriction item

Limit

Adjustable

Prefix lists per account per region

100

Not adjustable

Entries per prefix list

200

Not adjustable

Associated instances per prefix list

1,000

Not adjustable

For more information, see Prefix list overview.

Security groups

Restriction item

Basic security groups

Enterprise security groups

Maximum security groups per account per region

Quota ID: q_security-groups. See View or increase ECS quotas.

Same as basic security groups

Security groups per ENI

10

10

Maximum rules (inbound + outbound) across all security groups associated with a single ENI

1,000

1,000

Maximum rules per security group that reference another security group as the authorization object

20

0 — enterprise security groups do not support rules that use a security group as the authorization object, and cannot be used as the authorization object in other security group rules.

Instances per VPC security group

No fixed limit; determined by the number of private IP addresses the security group can contain.

No limit

Maximum private IP addresses per VPC security group (per account per region)

6,000. The count includes all private IP address types on associated ENIs: primary private IPv4, IPv6, secondary private IPv4, IPv4 prefixes, and IPv6 prefixes. If more than 6,000 private IP addresses need to communicate internally, distribute the instances across multiple security groups and configure rules to allow access between groups. View the limit using quota ID q_vpc-normal-security-group-ip-count in Quota Center.

65,536. The count includes all ENIs associated with the security group, including both primary and secondary network interfaces.

Port 25

Restricted by default for security reasons. Use SSL-encrypted port 465 to send emails instead.

Same as basic security groups

For more information, see Security group overview.

Reserved instances

Restriction item

Limit

Adjustable

Regional reserved instances per account (across all regions)

20

Submit a ticket

Zonal reserved instances per account per zone

20

Submit a ticket

Supported instance types

The gn6i and t5 instance families do not support regional reserved instances, splitting, or merging. Available instance types are shown on the purchase page.

Not adjustable

Resources that can be offset

Pay-as-you-go instances only (spot instances excluded). Computing resources (vCPUs and memory) only — network and storage costs cannot be offset. Windows-type reserved instances also offset image costs. See Billing overview.

Not adjustable

For more information, see What is a reserved instance?.

Savings plans

Restriction item

Limit

Adjustable

Savings plans per account

200

Not adjustable

Resources that can be offset

Pay-as-you-go ECS and Elastic Container Instance (ECI) instances (spot instances excluded). For ECS instances: computing resources (vCPUs and memory), images, system disks, data disks (capacity, provisioned performance, and burst fees), and fixed public bandwidth. For ECI instances with unspecified types: computing resources only. See Savings plan offset items and rules and Billing overview.

Not adjustable

For more information, see What is a savings plan?.

Storage Capacity Units

Restriction item

Limit

Adjustable

Maximum capacity per Storage Capacity Unit (SCU) purchase

50 TiB

Submit a ticket

Maximum SCUs per region

100

Not adjustable

Supported products

ESSDs, standard SSDs, ultra disks, and basic disks; Capacity and Performance Network Attached Storage (NAS) file systems; standard snapshots; Standard, Infrequent Access, and Archive OSS; Cloud Backup vault storage; Photo Album and Cloud Storage.

Not adjustable

Offset type

Pay-as-you-go bills only. Pay-as-you-go disk bills for spot instances cannot be offset.

Not adjustable

Effective time

Set when purchasing. Cannot be more than 6 months after the creation time.

Not adjustable

API management

Not supported. SCUs must be created and managed through the console.

Not adjustable

For more information, see Storage Capacity Unit (SCU).

Launch templates

Restriction item

Limit

Adjustable

Launch templates per account per region

Quota ID: q_launch-template-count. See View or increase ECS quotas.

View or increase ECS quotas

Versions per launch template per region

Quota ID: q_launch-template-version-count. See View or increase ECS quotas.

View or increase ECS quotas

Required parameters

All parameters are optional when creating a template. If required parameters (such as instance type or image) are omitted, provide them when using the template to create an instance.

Not adjustable

Modifying a launch template

Cannot be modified after creation. Create a new version to change configuration. See Manage launch template versions.

Not adjustable

For more information, see Launch template overview.

Deployment sets

Restriction item

Limit

Adjustable

Deployment sets per account per region

Quota ID: q_deployment-set-count. See View or increase ECS quotas.

View or increase ECS quotas

Instances per deployment set

Determined by the deployment strategy. See Deployment strategies.

Not adjustable

Dedicated hosts in a deployment set

Not supported.

Not adjustable

Region and zone requirements

The instance and its deployment set must be in the same region. For the low-latency network strategy, all instances must also be in the same zone.

Not adjustable

Supported instance types

Varies by deployment strategy. Call DescribeDeploymentSetSupportedInstanceTypeFamily with a deployment strategy to get the supported instance families.

Not adjustable

Merging deployment sets

Not supported.

Not adjustable

For more information, see Deployment sets.

Auto provisioning groups

Restriction item

Limit

Adjustable

Cross-region provisioning

Not supported.

Not adjustable

Configuration sources per group

One specific launch template version as the base configuration. Instance types in the template can be extended to form multiple resource pools.

Not adjustable

Resource pools per group

20 (each pool is a zone + instance type combination)

Not adjustable

Instances per group

1,000

Not adjustable

For more information, see Auto provisioning group overview.

Cloud Assistant

Restriction item

Limit

Adjustable

Cloud Assistant commands per account per region

Quota ID: q_axt-command-count. See View or increase ECS quotas.

View or increase ECS quotas

Task output size limit per account per region

Quota ID: q_axt-task-output-size. See View or increase ECS quotas.

Not adjustable

Task output retention period per account per region

Quota ID: q_axt-task-output-life. See View or increase ECS quotas.

Not adjustable

Activation codes for managed instances per account per region

Quota ID: q_cloud-assistant-activation-count. See View or increase ECS quotas.

Not adjustable

Instances per command execution per account per region

Quota ID: q_task-instance-count. See View or increase ECS quotas.

View or increase ECS quotas

Script size limits (Base64-encoded Bat, PowerShell, or Shell scripts, including custom parameters)

Create command: 18 KB after Base64 encoding. Run and save command: 18 KB. Run command without saving: 24 KB. Upload file: 32 KB.

Not adjustable

Custom parameters per command

20

Not adjustable

Supported operating systems

Alibaba Cloud Linux; CentOS 6, 7, 8, and later; CoreOS; Debian 8, 9, 10, and later; OpenSUSE; Red Hat Enterprise Linux 5, 6, 7, and later (requires manual RPM package installation — see Install the Cloud Assistant Agent); SUSE Linux Enterprise Server 11, 12, 15, and later; Ubuntu 12, 14, 16, 18, and later; FreeBSD 11, 12, 13, 14, and later; Windows Server 2012, 2016, 2019, and later.

Not adjustable

The Cloud Assistant Agent is pre-installed on instances created from ECS public images. For instances created from custom images or Alibaba Cloud Marketplace images, verify OS support and install the agent manually. See Install the Cloud Assistant Agent.

For more information, see Cloud Assistant overview.

Network connectivity diagnosis

Restriction item

Limit

Adjustable

Diagnostic lines per region

100

Not adjustable

Diagnostic tasks per region

1,000

Not adjustable

Concurrent diagnostic tasks per region

5

Not adjustable

For more information, see Diagnose network connectivity.

API rate limits

An API rate limit constrains the frequency of OpenAPI calls. ECS enforces two types of rate limits based on API version and resource type.

API type

Applies to

Adjustable

Elastic Compute Service API rate limit

APIs with version 2014-05-26 for resources such as images, security groups, and block storage. See View the ECS API rate limit.

Not adjustable

Elastic Block Storage API rate limit

Advanced feature APIs for block storage with version 2021-07-30. See View or increase the Elastic Block Storage API rate limit.

View or increase the Elastic Block Storage API rate limit

For more information about ECS APIs, see Integration overview.