All Products
Search
Document Center

ApsaraDB RDS:RDS PostgreSQL IO performance burst

Last Updated:Sep 24, 2025

If your business experiences significant fluctuations and frequent traffic peaks, you can enable the IO performance burst feature for premium performance disks. This feature allows your disk's input/output operations per second (IOPS) to exceed its maximum limit. This provides higher I/O capacity during peak hours to meet sudden business demands.

Background

Premium performance disks are a new storage class for RDS for PostgreSQL. They are compatible with the features of Enterprise SSDs (ESSDs). Premium performance disks also support the IO performance burst feature. This feature delivers higher IOPS performance under high I/O pressure without changing the storage capacity.

Feature overview

When you enable the IO performance burst feature for a premium performance disk, the maximum IOPS and throughput for the instance increase. The following table describes the differences.

Note

When you enable the IO performance burst feature on a primary node, the feature is also enabled on its secondary node. The feature is not automatically enabled for read-only instances. You must enable it on the details page of each read-only instance.

Scenario

Maximum IOPS

Maximum throughput (MB/s)

IO performance burst disabled

min{50000, max IOPS of instance type, 1800 + 50 × storage capacity}

min{350, max IO bandwidth of instance type, 120 + 0.5 × storage capacity}

IO performance burst enabled

min{1000000, max IOPS of instance type}

min{4000, max IO bandwidth of instance type}

Note
  • When throughput reaches its limit, IOPS is affected, and vice versa.

  • The preceding formulas do not apply to instances that use general-purpose instance types. General-purpose instance types use shared resources and cannot guarantee maximum IOPS or maximum bandwidth.

  • For the maximum IOPS and maximum bandwidth of the instance types in the preceding formulas, see RDS for PostgreSQL primary instance types.

Scenarios

A high I/O load can be caused by peak business hours, large transactions, full table scans, or queries that return large amounts of data. In these situations, CPU, memory, and storage capacity might be sufficient, but I/O utilization can approach or reach 100%. To handle the load, you might need to upgrade the instance configuration or the storage level, such as upgrading from a PL1 ESSD to a PL2 ESSD.

  • If you do not upgrade, an I/O load that exceeds the instance's I/O limit can cause performance degradation, query timeouts, or connection errors.

  • If you upgrade the instance configuration, I/O performance meets the demand, but this can lead to wasted CPU, memory, and storage resources.

  • If you upgrade the storage level, you can meet short-term I/O demands. However, when the I/O load decreases, the higher storage level becomes a waste of resources and money.

The IO performance burst feature for premium performance disks solves these problems. When enabled, the feature automatically triggers a performance burst during high I/O loads to increase the I/O limit. When the I/O load drops, the I/O limit returns to normal. This process provides serverless I/O performance and prevents wasted resources and costs.

Benefits

Compared to upgrading the instance configuration or storage level, the IO performance burst feature for premium performance disks offers several advantages that can significantly save costs.

  • You are billed only for the burst IOPS that you use.

  • Burst IOPS are calculated per second.

  • The feature includes a free quota. You are charged only for burst IOPS that exceed the free quota.

Billing

You will not be charged for using the IO performance burst feature before September 16, 2025 (Singapore time). Billing for this feature will officially start on September 16, 2025. For more information, see the official billing announcement.

Billing method

Premium performance disk fee = Storage fee + IO performance burst fee

  • Storage fee: This is the fee for the storage capacity of your RDS for PostgreSQL instance. It supports the subscription and pay-as-you-go billing methods. The billing method is the same as the method for PL1 ESSDs. For more information, see Billing items.

  • IO performance burst fee: Hourly IO performance burst fee = (Total burst IOPS of all nodes in the instance - Free quota) × Unit price of burst IOPS

    Note

    The IO performance burst fee for an RDS instance, including its primary and secondary nodes, is pay-as-you-go. You are charged if the IO performance burst feature is enabled and your usage exceeds the free quota.

    • Total burst IOPS of all nodes in the instance

      RDS Edition

      Instance burst IOPS

      Basic Edition

      Burst IOPS of the single node

      High-availability Edition

      Burst IOPS of primary node + Burst IOPS of secondary node

      Cluster Edition

      Burst IOPS of primary node + Burst IOPS of all secondary nodes

      Calculation method for single-node burst IOPS

      Scenario

      Burst IOPS amount

      Example

      IOPS exceeds the baseline IOPS, but throughput does not exceed the baseline throughput

      (IOPS - Baseline IOPS) × Duration

      Note

      Baseline IOPS refers to the maximum IOPS when the IO performance burst feature is disabled.

      The instance IOPS exceeds the baseline IOPS by 4,000 for 2 seconds. The burst IOPS amount is 4,000 × 2 = 8,000.

      Throughput exceeds the baseline throughput, but IOPS does not exceed the baseline IOPS

      (Throughput - Baseline throughput) × 1024 × Duration ÷ 16

      Note
      • Baseline throughput refers to the maximum throughput when the IO performance burst feature is disabled.

      • The unit of throughput is MB/s, and the unit of duration is seconds.

      16 is the conversion coefficient.

      The instance throughput exceeds the baseline throughput by 8 MB/s for 2 seconds. The burst IOPS amount is 8 × 1024 × 2 ÷ 16 = 1,024.

      Both IOPS and throughput exceed their baselines

      max{(IOPS - Baseline IOPS) × Duration, (Throughput - Baseline throughput) × 1024 × Duration ÷ 16}

      Note

      The unit of throughput is MB/s, and the unit of duration is seconds.

      If (IOPS - Baseline IOPS) × Duration is 8,000 and (Throughput - Baseline throughput) × 1024 × Duration ÷ 16 is 65,536, the final burst IOPS amount is 65,536.

      Note
      • When the primary node experiences high I/O pressure and generates a burst I/O, the secondary node also generates a corresponding burst I/O to ensure data and service consistency. The burst IOPS amount of the secondary node is almost the same as that of the primary node.

      • The IO performance burst feature for read-only instances must be enabled separately. The billing standard is the same as the standard for the primary instance. If a read-only instance is of the High-availability Edition, the fee includes the burst IOPS fees for its primary and secondary nodes.

    • Free quota:

      RDS provides a free quota for IO performance burst. When your burst IOPS usage exceeds the free quota, RDS for PostgreSQL charges a fee based on the instance edition.

      Product series

      Free quota

      Basic Edition

      300,000 IO/hour

      High-availability Edition

      600,000 IO/hour

      Cluster Edition

      800,000 IO/hour

    • Unit price of burst IOPS: USD 0.0015 per 10,000 IOPS

      Note

      If the burst IOPS usage that exceeds the free quota is less than 10,000, it is billed as 10,000 IOPS.

Billing example

Example scenario

An RDS for PostgreSQL instance has the following basic information: located in the China (Beijing) region, High-availability Edition, with 1,000 GB of storage. The baseline IOPS is 50,000. The burst rate is 20,000 IOPS. The burst lasts for 40 seconds every hour for a month.

The IO performance burst fee for this scenario is calculated as follows:

Burst IOPS amount

800,000 IO/hour

Free quota

600,000 IO/hour

Overage

200,000 IO/hour

Unit price

USD 0.0015 per 10,000 IO

Actual IO performance burst fee

0.0015 × (80 - 60) × 24 × 30 = USD 21.6

The following is a cost comparison:

Note

The prices in this example are for reference only. For actual prices, see the console.

Storage class

Storage unit price (USD/month)

Storage fee (USD)

IO performance burst fee (USD)

Total fee (USD/month)

Premium performance disk

244.8

244.8 × 1 = 244.8

21.6

244.8 + 21.6 = 266.4

PL2 ESSD

489.6

489.6 × 1 = 489.6

Not applicable

489.6

Monthly savings with a premium performance disk compared to a PL2 ESSD: 489.6 - 266.4 = USD 223.2

Note

Within a given period, the shorter the duration of high I/O on the instance, the more cost-effective a premium performance disk is compared to a PL2 ESSD.

Enable or disable IO performance burst

Note
  • Enabling or disabling the IO performance burst feature takes several minutes. The exact duration depends on the instance's usage, such as its read and write traffic.

  • This process does not cause transient connections and generally does not affect your business. In rare cases, the instance's IOPS might fluctuate. We recommend that you perform this operation during off-peak hours.

  1. Go to the Instances page. In the top navigation bar, select the region in which the RDS instance resides. Then, find the RDS instance and click the ID of the instance.

  2. In the Basic Information section, next to Storage Class, click Premium Performance Disk Switch Settings. In the dialog box that appears, enable or disable the IO Performance Burst switch.

    image

References

  • For more information about other storage classes supported by RDS for PostgreSQL, see Storage classes.

  • For more information about premium performance disks, see Premium performance disks.

  • To address a database disk I/O performance bottleneck, you can use the Buffer Pool Extension (BPE) feature of premium performance disks. This feature can significantly improve database I/O performance when you are facing large-scale or frequent data read and write demands. For more information, see Buffer Pool Extension (BPE).

  • To reduce storage costs, you can use the data archiving feature of premium performance disks. This feature uses Object Storage Service (OSS) as the storage medium to archive cold data, which significantly reduces storage costs. For more information, see Data archiving.

Related API operations

API

Description

ModifyDBInstanceSpec

When you configure the IO performance burst feature for a premium performance disk:

  • Make sure that the DBInstanceStorageType parameter is set to general_essd, which indicates that the instance uses a premium performance disk.

  • The BurstingEnabled parameter controls whether to enable or disable the IO performance burst feature. Set it to true to enable the feature or false to disable it.

  • Other parameters, such as the instance type and storage capacity, must be the same as the original parameter values of the instance. You cannot modify them.