A burstable iInstance is a cost-effective instance type for entry-level computing scenarios that experience occasional spikes in performance demand. This topic introduces the features and use cases of burstable instance families, explains concepts such as baseline performance, CPU credits, and performance modes, and lists the available instance types.
What is a burstable instance
A burstable instance is an instance type that uses CPU credits to govern its computing performance. It is suitable for workloads that typically have low CPU utilization but experience occasional, temporary spikes in demand. After you create a burstable instance, it continuously earns CPU credits. When the instance's performance is insufficient to handle its workload, it consumes these credits to boost computing performance without affecting the applications and environments on the instance. Burstable instances offer more flexible CPU usage at a lower cost than other instance types.
CPU credits allow you to shift computing capacity from off-peak to peak periods to save costs. If you experience unplanned high-performance demands, you can also enable the unlimited mode for your burstable instances.
Burstable instance families include the following:
A burstable instance is a specific type of shared instance. For information about other shared instance families, see Shared instance families.
The following table describes the baseline performance, CPU credits, and performance modes of burstable instances.
Term | Description | References |
Baseline performance | The sustained CPU performance that an instance can deliver, which is determined by its instance type. | |
Initial CPU credits | A one-time grant of CPU credits an instance receives upon creation, fixed at 30 credits per vCPU. | |
CPU credit balance | The accumulated CPU credits that are stored when an instance earns more credits than it consumes. These credits are used to boost CPU utilization above the baseline performance. | |
Max CPU credit balance | The total number of CPU Credits an instance can accumulate. This is equivalent to the credits earned in a 24-hour period. Unused credits expire after 24 hours, so the balance is capped at this maximum and maintained in a dynamic equilibrium. | |
Performance mode | Two modes: standard and unlimited.
| |
Advance CPU credits | Credits earned over the next 24 hours. May incur extra charges. Only available in unlimited mode. | |
Overdrawn CPU credits | Credits consumed after depleting advance credits to maintain CPU utilization above baseline. Incurs extra charges. Only available in unlimited mode. |
Use cases
When you purchase an enterprise-level instance, you gain full access to its vCPUs. This means you pay for the entire vCPU capacity, whether your CPU utilization is 0% or 100%. If your workload is predictable and only requires high CPU performance during specific periods, you are paying for unused computing resources during the rest of the time. In this situation, you can choose a burstable instance to save costs while meeting performance needs.
Burstable instances are ideal for workloads that require sudden bursts of computing performance, such as development and testing environments, lightweight applications, microservices, and web application servers. Before purchasing, evaluate your workload's performance needs during both off-peak and peak periods. To save costs while meeting overall performance demands, select an instance type with a baseline performance that at least meets your off-peak requirements.
If you find that your selected burstable instance type does not meet your needs, you can change its configuration. For more information, see Instance type changes.
Windows applications and graphical user interfaces (GUIs) require high CPU resources. Running these workloads on a t-series burstable instance may cause system freezes or crashes. We recommend you select an appropriate instance type based on your actual needs, such as a General-purpose (g-series), Compute-optimized (c-series), or Memory-optimized (r-series) instance type.
Baseline performance
An instance's type determines its baseline performance. It represents the sustained CPU performance the instance provides. You can find the baseline performance for different instance types in the Average baseline CPU performance column of the instance specification tables.
CPU credits
CPU credits represent the computing capacity you hold and determine the performance a burstable instance can achieve. The related concepts and examples are as follows.
Initial CPU credits
To ensure you have CPU credits available for deployment immediately after startup, each vCPU of a new burstable instance receives 30 initial CPU credits.
For example, an ecs.t5-lc1m2.large instance with 2 vCPUs receives 60 Initial CPU credits upon creation. An ecs.t5-c1m1.xlarge instance with 4 vCPUs receives 120 Initial CPU credits upon creation.
Credit earning rate
After an instance starts, it begins consuming CPU credits to maintain its performance while simultaneously earning credits at a fixed rate. This earning rate is determined by the instance type and can be found in the CPU credits/hour column of the instance specification tables. This metric indicates the total credits that all vCPUs of a single instance can earn per hour.
For example, an ecs.t5-c1m1.large instance has a baseline performance of 25%. This means the CPU credits it earns per hour can sustain one vCPU at 25% utilization for one hour, or at 100% utilization for 15 minutes (60 minutes × 25%). Corresponding to its baseline performance, each vCPU earns 15 CPU credits per hour. Since the ecs.t5-c1m1.large instance has 2 vCPUs, it earns a total of 30 CPU credits per hour.
CPU credit balance
If an instance earns more CPU credits than it consumes, the surplus is stored as the CPU credit balance. These credits are held for a maximum of 24 hours, maintaining a dynamic equilibrium. Since the credit earning rate is fixed for each instance type, the CPU credit Balance has a maximum limit. This limit is equal to the total number of credits the instance can earn in 24 hours. You can find this value in the Max CPU credit balance column of the instance specification tables.
For example, an ecs.t5-c1m1.large instance earns 30 CPU credits per hour, so its maximum CPU credit balance is 720 (30 × 24).
Credit consumption
The rate at which CPU credits are consumed depends on the number of vCPUs, CPU utilization, and operating time of the burstable instance. For example, each of the following scenarios consumes 1 CPU credit:
1 vCPU runs at 100% utilization for 1 minute.
1 vCPU runs at 50% utilization for 2 minutes.
2 vCPUs run at 25% utilization for 2 minutes.
A burstable instance begins consuming CPU credits to maintain performance as soon as it starts. It first consumes the Initial CPU credits. Once depleted, Initial CPU credits do not replenish. After that, the instance can only consume the credits it earns.
When CPU utilization is below the baseline performance, the instance consumes fewer credits than it earns, and the CPU credit balance increases.
When CPU utilization is equal to the baseline performance, the instance consumes the same number of credits as it earns, and the CPU credit balance remains unchanged.
When CPU utilization is above the baseline performance, the instance consumes more credits than it earns, and the CPU credit balance decreases.
CPU utilization for credit consumption is measured at the physical host level and includes the overhead for emulating privileged instructions within the virtual machine. You can view this data in the CloudMonitor console. On the Host monitoring page, click an instance ID, and then view the data on the Basic monitoring tab. For more information, see Host monitoring overview.
How instance stops affect CPU credit earning
Pay-as-you-go instance stopped (standard mode): The current credit balance is retained, and the instance continues to earn CPU credits.
Pay-as-you-go instance stopped (economical mode): The current credit balance expires, and the instance stops earning CPU credits. When restarted, the instance receives new Initial CPU credits and resumes earning credits.
Pay-as-you-go instance stopped due to an overdue payment: The current credit balance is retained, but the instance stops earning CPU credits. It will resume earning credits after the bill is paid.
Subscription instance stopped after expiration: The current credit balance is retained, but the instance stops earning credits. It will resume earning credits when restarted.
Subscription instance stopped before expiration: The current credit balance is retained, and the instance continues to earn CPU credits.
Performance modes
Burstable instances can run in either standard mode or unlimited mode.
Standard mode
In standard mode, the performance of a burstable instance is constrained by its CPU credits. Once the initial CPU credits and CPU credit balance are depleted, the instance's performance cannot exceed its baseline performance. However, to prevent a sharp performance drop, the instance's performance will gradually decrease to the baseline level over a 15-minute period as the credit balance runs low.
Standard mode is suitable for workloads with stable and predictable CPU usage that rarely exceeds the baseline performance but occasionally requires a performance boost. Examples include lightweight web servers, development and test environments, and low-to-medium performance databases.
Unlimited mode
In unlimited mode, a burstable instance can exceed its available CPU credits. It can maintain CPU utilization above its baseline performance at any time using advance CPU credits or overdrawn CPU credits. After the initial CPU credits and CPU credit balance are depleted, if the instance's CPU utilization remains above its baseline performance, it begins to consume advance CPU credits and then overdrawn CPU credits.
Advance CPU credits: Credits earned over the next 24 hours. May incur extra charges.
Overdrawn CPU credits: Credits consumed after depleting advance credits to stay above baseline. Always incurs extra charges.
For more information about extra charges for advance and overdrawn credits, see Extra charges.
The following diagram shows CPU credit changes in unlimited mode.
If your instance consumes advance CPU credits and is then stopped (in economical mode), undergoes a configuration change, is released, or is switched to standard mode before the advance CPU credits are fully replenished, you will be billed a one-time fee for the used credits.
Unlimited mode is suitable for workloads with sudden performance spikes that the CPU credit balance cannot cover, potentially requiring the use of Advance CPU credits or even overdrawn CPU credits. Examples include:
Scenarios where high traffic is anticipated and high performance must be maintained for a specific period, such as a product launch, an e-commerce promotion, or a website hosting a marketing campaign. You can temporarily enable unlimited mode and then disable it after the peak period to reduce costs.
Scenarios where a website's workload peaks at specific times, but its average 24-hour CPU utilization remains below the baseline performance. You can keep unlimited mode enabled to ensure a smooth user experience during peak hours. If the credits earned during low-traffic periods are sufficient to replenish the consumed advance CPU credits, you can maintain a good overall user experience without incurring additional charges.
By default, a burstable instance is created in standard mode. If you want to use unlimited mode, see Enable unlimited mode.
For examples of how CPU credits change in different performance modes, see CPU credit change examples.
Instance configuration changes
If monitoring shows that your burstable instance's CPU utilization is consistently higher or lower than its baseline performance, the current instance type may not be the right fit for your workload. We recommend you re-evaluate whether the current instance type is appropriate and change it if necessary. The procedure for an instance configuration change depends on the billing method. For more information, see Instance configuration changes overview.
t6, burstable instance family
Features of t6:
vCPUs deliver steady baseline performance with the capability to burst, which is limited by CPU credits.
Provides a better cost-performance ratio than the previous-generation t5 family.
Compute:
Processor: 2.5 GHz base clock speed and 3.2 GHz turbo frequency latest-generation Intel® Xeon® Cascade Lake server processors
DDR4 memory
Storage:
I/O optimized
Supports Enterprise SSDs (ESSDs), ESSD AutoPL disks, standard SSDs, and ultra disks.
ImportantDue to the limits of burstable instances, ESSDs at performance level (PL) 2 and PL3 cannot deliver their peak performance. We recommend that you use enterprise-grade instance types or ESSDs at lower PLs instead.
Network:
Supports IPv4 and IPv6
Supports only virtual private clouds (VPCs)
Use cases:
Web application servers
Lightweight applications and microservices
Development, testing, and stress testing environments
t6 instance types
Instance type | vCPU | Memory (GiB) | Average baseline CPU performance | CPU credits/hour | Max CPU credit balance | Network baseline/burst bandwidth (Gbit/s) | Packet forwarding rate (pps) | Multi-Queue | ENIs | Private IPv4 addresses per ENI | IPv6 addresses per ENI |
ecs.t6-c4m1.large | 2 | 0.5 | 5% | 6 | 144 | 0.08/up to 0.4 | 40,000 | 1 | 2 | 2 | 1 |
ecs.t6-c2m1.large | 2 | 1.0 | 10% | 12 | 288 | 0.08/up to 0.6 | 60,000 | 1 | 2 | 2 | 1 |
ecs.t6-c1m1.large | 2 | 2.0 | 20% | 24 | 576 | 0.08/up to 1 | 100,000 | 1 | 2 | 2 | 1 |
ecs.t6-c1m2.large | 2 | 4.0 | 20% | 24 | 576 | 0.08/up to 1 | 100,000 | 1 | 2 | 2 | 1 |
ecs.t6-c1m4.large | 2 | 8.0 | 30% | 36 | 864 | 0.08/up to 1 | 100,000 | 1 | 2 | 2 | 1 |
ecs.t6-c1m4.xlarge | 4 | 16.0 | 40% | 96 | 2304 | 0.16/up to 2 | 200,000 | 1 | 2 | 6 | 1 |
ecs.t6-c1m4.2xlarge | 8 | 32.0 | 40% | 192 | 4608 | 0.32/up to 4 | 400,000 | 2 | 2 | 6 | 1 |
This instance family does not support adding secondary elastic network interfaces (ENIs) during instance creation. You can attach one after the instance is created. For some instance types, the instance must be in the Stopped state to attach or detach a secondary ENI. These types include
ecs.t6-c1m1.large,ecs.t6-c1m2.large,ecs.t6-c1m4.large,ecs.t6-c2m1.large, andecs.t6-c4m1.large.You can go to the Instance Types Available for Each Region page to view the instance types available in each region.
For more information about metric definitions, see Instance type metrics.
t5, burstable instance family
Features of t5:
vCPUs deliver steady baseline performance with the capability to burst, which is limited by CPU credits.
Balanced compute, memory, and network resources
Compute:
Provides multiple processor and memory configurations.
Processor: 2.5 GHz base clock speed Intel® Xeon® processors
DDR4 memory
Storage: Supports only ultra disks and standard SSDs
Network:
Supports IPv4 and IPv6
Supports only VPCs
Use cases:
Web application servers
Lightweight applications and microservices
Development, testing, and stress testing environments
t5 instance types
Instance type | vCPU | Memory (GiB) | Average baseline CPU performance | CPU credits/hour | Max CPU credit balance | Network baseline bandwidth (Gbit/s) | Packet forwarding rate (pps) | Multi-Queue | ENIs | Private IPv4 addresses per ENI | IPv6 addresses per ENI |
ecs.t5-lc2m1.nano | 1 | 0.5 | 20% | 12 | 288 | 0.1 | 40,000 | 1 | 2 | 2 | 1 |
ecs.t5-lc1m1.small | 1 | 1.0 | 20% | 12 | 288 | 0.2 | 60,000 | 1 | 2 | 2 | 1 |
ecs.t5-lc1m2.small | 1 | 2.0 | 20% | 12 | 288 | 0.2 | 60,000 | 1 | 2 | 2 | 1 |
ecs.t5-lc1m2.large | 2 | 4.0 | 20% | 24 | 576 | 0.4 | 100,000 | 1 | 2 | 2 | 1 |
ecs.t5-lc1m4.large | 2 | 8.0 | 20% | 24 | 576 | 0.4 | 100,000 | 1 | 2 | 2 | 1 |
ecs.t5-c1m1.large | 2 | 2.0 | 25% | 30 | 720 | 0.5 | 100,000 | 1 | 2 | 2 | 1 |
ecs.t5-c1m2.large | 2 | 4.0 | 25% | 30 | 720 | 0.5 | 100,000 | 1 | 2 | 2 | 1 |
ecs.t5-c1m4.large | 2 | 8.0 | 25% | 30 | 720 | 0.5 | 100,000 | 1 | 2 | 2 | 1 |
ecs.t5-c1m1.xlarge | 4 | 4.0 | 25% | 60 | 1440 | 0.8 | 200,000 | 1 | 2 | 6 | 1 |
ecs.t5-c1m2.xlarge | 4 | 8.0 | 25% | 60 | 1440 | 0.8 | 200,000 | 1 | 2 | 6 | 1 |
ecs.t5-c1m4.xlarge | 4 | 16.0 | 25% | 60 | 1440 | 0.8 | 200,000 | 1 | 2 | 6 | 1 |
ecs.t5-c1m1.2xlarge | 8 | 8.0 | 25% | 120 | 2880 | 1.2 | 400,000 | 1 | 2 | 6 | 1 |
ecs.t5-c1m2.2xlarge | 8 | 16.0 | 25% | 120 | 2880 | 1.2 | 400,000 | 1 | 2 | 6 | 1 |
ecs.t5-c1m4.2xlarge | 8 | 32.0 | 25% | 120 | 2880 | 1.2 | 400,000 | 1 | 2 | 6 | 1 |
ecs.t5-c1m1.4xlarge | 16 | 16.0 | 25% | 240 | 5760 | 1.2 | 600,000 | 2 | 2 | 6 | 1 |
ecs.t5-c1m2.4xlarge | 16 | 32.0 | 25% | 240 | 5760 | 1.2 | 600,000 | 2 | 2 | 6 | 1 |
This instance family does not support adding secondary ENIs during instance creation. You can attach one after the instance is created. For some instance types, the instance must be in the Stopped state to attach or detach a secondary ENI. These types include
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, andecs.t5-lc1m4.large.You can go to the Instance Types Available for Each Region page to view the instance types available in each region.
For more information about metric definitions, see Instance type metrics.