This topic describes the features, specific instance types, baseline performance, CPU credits, and performance modes of burstable instances. You can use CPU credits to burst the performance of a burstable instance when required by your workloads.

Scenarios

After you purchase an enterprise-level instance, the vCPUs of the instance are reserved only for you. During this period, you are charged for the vCPU resources regardless of whether you fully utilize the performance of the vCPUs. Even if you require high levels of computing power only for a specific period during a day, you are still charged for the unused resources for the whole day. To avoid this situation, you can use burstable instances to better meet your business requirements.

Burstable instances apply to scenarios where you require higher-than-normal performance for a specific period, such as stress testing service applications, lightweight applications, microservices, and web application servers. We recommend that you evaluate your business requirements to determine the performance levels required during off-peak and peak hours before you make a purchase. The baseline performance of the instances that you purchase must meet your business requirements during off-peak hours. This way, you can enjoy the required performance at a significantly lower cost.
Note If the selected burstable instances cannot meet your requirements, you can change the configurations. For more information, see Change configurations.

Features

Burstable instances are special shared instances. For information about other shared instance families, see Shared instance families.

You can accumulate CPU credits for your burstable instances, and consume these credits to increase the computing performance of instances when required by your workloads. This consumption pattern does not affect the environments or applications that are running on your instances. For more information about the baseline performance and CPU credits of burstable instances, see Baseline performance, Earn CPU credits, and Consume CPU credits.

The CPU credit mechanism allows you to minimize consumption of resources during off-peak hours, and scale resources during peak hours at no extra cost. If you have unplanned performance requirements, you can enable the unlimited mode for your instances.

An instance in standard mode will run below its baseline performance if its CPU credits are depleted. However, a burstable instance in unlimited mode allows you to overdraw or pay for additional CPU credits and utilize the CPU above its baseline performance at any time. In this case, you may be charged for the consumption of these additional CPU credits. For more information about the performance modes, see Standard mode and Unlimited mode.

Baseline performance

Baseline performance is the amount of vCPU capacity that is continuously provisioned to a burstable instance. Baseline performance varies depending on the instance type. If the vCPUs of your burstable instances are running at the baseline performance, the CPU credits that you earn are equal to the CPU credits that you consume. The Baseline CPU computing performance column in the burstable instance family table indicates the baseline performance of a single vCPU.

Earn CPU credits

CPU credits can be described as computing resources that are available. These computing resources determine the computing performance of your burstable instances.

After you create a burstable instance, 30 CPU credits are provisioned for each vCPU of the instance, which are initial CPU credits. These credits enable you to complete deployment tasks after you start the instance. For example, an ecs.t5-lc1m2.large instance has two vCPUs. You will earn 60 initial CPU credits after you create one instance of this type. An ecs.t5-c1m1.xlarge instance has four vCPUs. You will earn 120 initial CPU credits after you create one instance of this type.

After you start a burstable instance, the instance starts to consume CPU credits to maintain its computing performance. At the same time, the instance also earns CPU credits at a fixed rate that is determined by the instance type. The CPU credits that a vCPU can earn per hour are based on its baseline performance. The CPU credits per hour column in the t5 instance type table indicates the CPU credits that all of the vCPUs of an instance can earn per hour. For example, 25% baseline performance of an ecs.t5-c1m1.large instance indicates that the CPU credits that a vCPU of the instance earns per hour can keep the vCPU running at 25% utilization for an hour or at 100% utilization for 15 minutes (60 × 25%). In response to its baseline performance, each vCPU earns 15 CPU credits per hour. Therefore, an ecs.t5-c1m1.large instance with two vCPUs earns 30 CPU credits per hour.

If the CPU credits earned exceed the credits consumed, the net credits are accrued as CPU credit balance. Your CPU credit balance is valid for 24 hours to ensure the availability of CPU credits. A specific instance type earns CPU credits at a fixed rate. Therefore, its CPU credit balance is limited. The maximum CPU credit balance of a specific instance type is the number of CPU credits that the instance can earn within 24 hours. For more information, see the Max CPU credit balance column in the burstable instance family table. For example, an ecs.t5-c1m1.large instance can earn 30 CPU credits per hour. This way, the maximum CPU credit balance that the instance can earn is 720 (30 × 24).

In different scenarios, the shutdown of your instances may affect CPU credits:
  • If the No fees for Stopped Instances (VPC-Connected) feature is disabled, your current CPU credit balance is retained and the instances continue to earn CPU credits.
  • If the No fees for Stopped Instances (VPC-Connected) feature is enabled, your current CPU credit balance becomes invalid and the instances cannot continue to earn credits. After you restart your instances, you receive initial credits and start to earn credits again.
  • When a subscription instance expires, your current CPU credit balance is retained, but you cannot continue to earn credits. When you restart the instance, you start to earn credits again.
  • If your payment for a pay-as-you-go instance is overdue, your current CPU credit balance is retained, but you cannot continue to earn credits until you complete the payment.

Consume CPU credits

The consumption rate of CPU credits of a burstable instance is based on the number of vCPUs, CPU utilization, and operating hours. For example, you consume one CPU credit in the following scenarios:
  • A vCPU runs at 100% utilization for a minute.
  • A vCPU runs at 50% utilization for two minutes.
  • Two vCPUs run at 25% utilization for two minutes.
When you start a burstable instance, the instance starts to consume CPU credits to maintain its computing performance. Initial credits that cannot be replenished are consumed first. When the initial credits are used up, the instance consumes the accumulated CPU credits.
  • When your vCPUs run below the baseline performance, the earned credits are greater than the consumed credits. In this case, the CPU credit balance increases.
  • When your vCPUs run at the baseline performance, the earned credits are equal to the consumed credits. In this case, the CPU credit balance remains unchanged.
  • When your vCPUs run above the baseline performance, the earned credits are less than the consumed credits. In this case, the CPU credit balance decreases.

Standard mode

The performance of a burstable instance in standard mode is based on the availability of CPU credits. After the instance consumes all the initial credits and accrued CPU credits, the instance cannot run above the baseline performance. When the CPU credit balance is low, the instance gradually reduces performance to the baseline level within 15 minutes. This way, the instance does not experience a sharp performance drop-off when its accrued CPU credit balance is depleted.
Note For more information about the relationship between the performance and CPU credits of a burstable instance in standard mode, see CPU credit change examples.

The standard mode applies to scenarios where you have stable workloads, do not need instances to run above the baseline performance for an extended period of time, and may occasionally need burst performance, such as lightweight web servers, development and testing environments, and databases with low and medium performance.

Unlimited mode

The performance of a burstable instance in unlimited mode is not limited by the availability of CPU credits. You can overdraw or pay for additional CPU credits to obtain performance boosts at any time. If you keep running your instances above the baseline performance after the initial CPU credits and accrued credits are consumed, the CPU credits will change as shown in the following figure. The following section describes the two concepts involved in the figure:
  • Advance CPU credits: Advance CPU credits are credits that you can pay for in advance and can receive within 24 hours after you make the payment.
  • Overdrawn CPU credits: When you have consumed all the advance CPU credits, overdrawn CPU credits are used to make sure that the instance is running above the baseline performance. You will be charged for using overdrawn CPU credits.
Note For more information about the billing of burstable instances, see Impact of performance modes on billing.
Advance CPU credits and overdrawn CPU credits
Note For more information about the relationship between the performance and CPU credits of an instance in unlimited mode, see CPU credit change examples.
You can enable the unlimited mode for your instances if you need to consume advance CPU credits or overdrawn CPU credits in addition to your credit balance to meet burstable performance requirements. Examples:
  • Some events, such as new feature releases, e-commerce promotions, and website promotions, will cause a substantial increase to your workloads. High CPU performance is required during this period of time. In this case, you can enable the unlimited mode for your instances. You can disable the mode to save costs when the workload peak ends.
  • Some web applications may require CPU bursts for a specific period during a day, but the daily average CPU utilization is below the baseline CPU utilization. In this case, you can enable the unlimited mode for your instances during peak hours to ensure a positive customer experience. If the CPU credits that you have earned during off-peak hours can offset the advance CPU credits that you consumed during peak hours, you can ensure overall positive customer experience without the need to pay extra charges.
Note By default, when you create a burstable instance, the standard mode is enabled. For more information about how to enable the unlimited mode, see Enable the unlimited mode.

Change configurations

During monitoring, you may find that the vCPUs are constantly running above or below the baseline performance. This indicates that the instance type cannot meet your business requirements. We recommend that you re-evaluate whether the instance type is suitable and select another burstable instance type or enterprise-level instance based on your needs. For more information, see Instance families that support instance type changes.

The operation of changing configurations varies based on the billing method:

t6, burstable instance family

Features:
  • Equipped with 2.5 GHz Intel ® Xeon ®Cascade Lake processors with a maximum turbo frequency of 3.2 GHz.
  • More cost-effective when compared with the t5 burstable instance family.
  • Delivers a bandwidth of up to 6 Gbit/s.
  • Paired with DDR4 memory.
  • Provides baseline CPU performance and is burstable, but limited by accumulated CPU credits.
  • Supports VPCs only.
  • Scenarios:
    • Web application servers
    • Lightweight applications and microservices
    • Development and testing environments
Instance types
Instance type vCPU Memory (GiB) Baseline CPU computing performance CPU credits per hour Max CPU credit balance Local storage (GiB) Base bandwidth (Gbit/s) Packet forwarding rate (Kpps) IPv6 support NIC queues ENIs (including one primary ENI) Private IP addresses per ENI
ecs.t6-c4m1.large 2 0.5 5% 6 144 None 0.08 40 Yes 1 2 2
ecs.t6-c2m1.large 2 1.0 10% 12 288 None 0.08 60 Yes 1 2 2
ecs.t6-c1m1.large 2 2.0 20% 24 576 None 0.08 100 Yes 1 2 2
ecs.t6-c1m2.large 2 4.0 20% 24 576 None 0.08 100 Yes 1 2 2
ecs.t6-c1m4.large 2 8.0 30% 36 864 None 0.08 100 Yes 1 2 2
ecs.t6-c1m4.xlarge 4 16.0 40% 96 2304 None 0.16 200 Yes 1 2 6
ecs.t6-c1m4.2xlarge 8 32.0 40% 192 4608 None 0.32 400 Yes 1 2 6
Note
  • When you bind or unbind an ENI, instances of the following instance types must be in the stopped state: ecs.t6-c1m1.large, ecs.t6-c1m2.large, ecs.t6-c1m4.large, ecs.t6-c2m1.large, and ecs.t6-c4m1.large.
  • You can go to the ECS Instance Types Available for Each Region page to view the instance types available in each region.
  • For more information about these specifications, see Instance specifications.

t5, burstable instance family

Features:
  • Equipped with 2.5 GHz Intel ® Xeon ® processors.
  • Paired with DDR4 memory.
  • Supports multiple CPU-to-memory ratios.
  • Provides baseline CPU performance and is burstable, but limited by accumulated CPU credits.
  • Offers a balance of compute, memory, and network resources.
  • Supports VPCs only.
  • Scenarios:
    • Web application servers
    • Lightweight applications and microservices
    • Development and testing environments
Instance types
Instance type vCPU Memory (GiB) Baseline CPU computing performance CPU credits per hour Max CPU credit balance Local storage (GiB) Bandwidth (Gbit/s) Packet forwarding rate (Kpps) IPv6 support NIC queues ENIs (including one primary ENI) Private IP addresses per ENI
ecs.t5-lc2m1.nano 1 0.5 20% 12 288 None 0.1 40 Yes 1 2 2
ecs.t5-lc1m1.small 1 1.0 20% 12 288 None 0.2 60 Yes 1 2 2
ecs.t5-lc1m2.small 1 2.0 20% 12 288 None 0.2 60 Yes 1 2 2
ecs.t5-lc1m2.large 2 4.0 20% 24 576 None 0.4 100 Yes 1 2 2
ecs.t5-lc1m4.large 2 8.0 20% 24 576 None 0.4 100 Yes 1 2 2
ecs.t5-c1m1.large 2 2.0 25% 30 720 None 0.5 100 Yes 1 2 2
ecs.t5-c1m2.large 2 4.0 25% 30 720 None 0.5 100 Yes 1 2 2
ecs.t5-c1m4.large 2 8.0 25% 30 720 None 0.5 100 Yes 1 2 2
ecs.t5-c1m1.xlarge 4 4.0 25% 60 1440 None 0.8 200 Yes 1 2 6
ecs.t5-c1m2.xlarge 4 8.0 25% 60 1440 None 0.8 200 Yes 1 2 6
ecs.t5-c1m4.xlarge 4 16.0 25% 60 1440 None 0.8 200 Yes 1 2 6
ecs.t5-c1m1.2xlarge 8 8.0 25% 120 2880 None 1.2 400 Yes 1 2 6
ecs.t5-c1m2.2xlarge 8 16.0 25% 120 2880 None 1.2 400 Yes 1 2 6
ecs.t5-c1m4.2xlarge 8 32.0 25% 120 2880 None 1.2 400 Yes 1 2 6
ecs.t5-c1m1.4xlarge 16 16.0 25% 240 5760 None 1.2 600 Yes 1 2 6
ecs.t5-c1m2.4xlarge 16 32.0 25% 240 5760 None 1.2 600 Yes 1 2 6
Note
  • When you bind or unbind an ENI, instances of the following instance types must be in the stopped state: ecs.t5-lc2m1.nano, ecs.t5-c1m1.large, ecs.t5-c1m2.large, ecs.t5-c1m4.large, ecs.t5-lc1m1.small, ecs.t5-lc1m2.large, ecs.t5-lc1m2.small, and ecs.t5-lc1m4.large.
  • You can go to the ECS Instance Types Available for Each Region page to view the instance types available in each region.
  • For more information about these specifications, see Instance specifications.

Terms of burstable instances

Term Description Reference
initial CPU credits The CPU credits that are provisioned when you create a burstable instance. You can earn 30 initial credits for each vCPU. These credits cannot be replenished when used up. For more information, see Earn CPU credits.
CPU credit balance The net credits that are accrued when the earned CPU credits exceed the consumed credits. You can use these credits to run instances above the baseline performance. For more information, see Earn CPU credits.
max CPU credit balance The CPU credits that a burstable instance can earn within 24 hours. Your CPU credit balance is valid for 24 hours to ensure the availability of CPU credits. A specific instance type earns CPU credits at a fixed rate. Therefore, its CPU credit balance is limited. For more information, see Earn CPU credits.
advance CPU credits The CPU credits that you can pay for in advance and can receive within 24 hours after you make the payment. You can use advance CPU credits only when you enable the unlimited mode. For more information, see Unlimited mode.
overdrawn CPU credits The CPU credits that you use when you have consumed all advance CPU credits to make sure that the instance is running above the baseline performance. You will be charged for using these credits. You can use overdrawn CPU credits only when you enable the unlimited mode. For more information, see Unlimited mode.