All Products
Search

This Product

    PolarDB:Temporarily add compute nodes

    Document Center

    PolarDB:Temporarily add compute nodes

    Last Updated:Jun 20, 2026

    When your business—such as E-commerce sales promotions, data analytics, or report generation—faces a predictable short-term spike in read traffic, insufficient database read capacity may slow down application response time. To handle this scenario, PolarDB offers the ability to temporarily add read-only nodes. In subscription clusters, you can add one or more read-only nodes on demand and set an automatic revert time. This feature provides temporary scalability for your cluster’s read processing capacity during traffic peaks. At the specified time, the system automatically removes the temporary nodes and restores the original cluster configuration without manual intervention.

    Impact assessment

    Before adding nodes temporarily, assess their potential impact on your business to plan the change effectively.

    Service impact

    • Add temporary nodes: Adding one temporary read-only node takes about 5 minutes. When adding multiple temporary read-only nodes simultaneously, PolarDB processes them in parallel. The actual time depends on factors such as cluster load and the number of databases and tables. The cluster remains unaffected during this process.

    • Remove temporary nodes: At the scheduled revert time, connections to the removed node experience a transient disconnection, but other nodes remain unaffected. If your application connects to the cluster endpoint, PolarDB automatically excludes the deleted node, so you do not need to update your application configuration.

    Node count limits

    • Enterprise Edition:

      • Cluster Edition: Supports up to 15 read-only nodes. You must keep at least one read-only node to ensure high availability. Therefore, you cannot delete the last read-only node.

      • Multi-master Cluster (Limitless): Supports up to 63 read/write nodes and 15 read-only nodes (global read-only nodes).

    • Standard Edition: Supports up to 7 read-only nodes.

    Node specification limits

    To ensure cluster stability and high availability, the following restrictions apply to combinations of primary and read-only node specifications:

    • The memory of a new read-only node must be at least half the memory of the primary node.

    • New read-only nodes must meet minimum CPU core requirements. The rules are as follows:

      Primary node CPU cores

      Minimum allowed CPU cores for read-only nodes

      2

      2

      4

      4

      8

      4

      16

      8

      32

      16

      64

      32

      88

      64

      120

      64
    Note

    The Multi-master Cluster (Limitless) series is not subject to these restrictions.

    Other limitations

    Applicability

    Before proceeding, complete the following checks.

    • Billing type: The destination cluster must use the Subscription billing model.

    • Cluster nodes:

      • Temporarily add read-only nodes:

        • The cluster must already have at least one read-only node. Single-node clusters do not support temporary node addition.

        • Multi-master Cluster (Limitless): Does not support adding temporary read-only nodes.

      • Temporarily add columnstore index read-only nodes:

        • The cluster must already have at least one columnstore index read-only node.

    • Cluster status: The destination cluster must be Running, with no ongoing configuration changes (for example, adding or deleting nodes, or minor version upgrades).

    • Configuration change task check: No pending configuration change tasks (for example, adding or deleting nodes, or changing configuration) should exist.

    Procedure

    1. Log on to the PolarDB console. In the navigation pane on the left, click Clusters and select the cluster’s region.

    2. Use one of the following methods to open the Add/Remove Node:

      • Find the target cluster in the cluster list and click Add/Remove Node in its Actions column.

      • Click the target cluster ID to go to its Basic Information page. In the Database Nodes section, click Add/Remove Node.

    3. In the Add/Remove Node dialog box, select either Temporarily Add Node or Temporarily Add Columnstore Index Read-Only Node based on your business needs, then click OK.

    4. On the Temporary Upgrade page, confirm the Current Configuration and Expired At time. Click Add Node or Add a Read-only IMCI Node and select the node specifications.

      Note

      You can click this button multiple times to add several nodes at once.

    5. Set the Restore Point: Specify a precise time when the system will automatically restore the cluster to its pre-upgrade configuration.

      • The minimum duration is 1 hour. The maximum is 14 days. For needs beyond 14 days, use manual configuration change.

      • The revert time must be at least one day before the cluster’s expiration date. For example, if the cluster expires on January 10, the latest revert time is January 9.

      • You cannot modify the revert time after setting it. If performance is still insufficient and you need to extend the duration, you can perform one additional temporary node addition before the revert time arrives. The Restore Point for this second operation cannot be earlier than the first.

    6. After reviewing the service agreement, click Buy Now and complete payment. After successful payment, the temporary node addition task starts.

    Billing details

    The cost of temporarily adding nodes is based on the price difference between the original and new specifications, multiplied by a 1.5 premium coefficient. The formula is:

    Cost per compute node for N days = (Monthly price of the new specification ÷ 30) × 1.5 × N.

    Related operations

    Convert temporary nodes to permanent nodes

    If you want to keep the added nodes permanently or need to perform operations such as manual configuration change, temporary specification upgrade, or cluster opt-out, convert the temporary nodes to permanent nodes (they will not be automatically deleted at the revert time) using one of the following methods.

    • Manual conversion: During the temporary node addition period, return to the Add or Remove Nodes wizard and perform an Add Nodes operation.

      Note

      This action creates only a permanent billing order and does not cause a second service interruption.

    • Conversion via storage expansion: If you manually scale out the cluster’s storage space during the temporary node addition period, the system automatically converts the temporary nodes to permanent nodes. From the moment of storage expansion, billing switches to the new subscription price. Carefully evaluate this change before expanding storage.

    View the expiration time of temporary node addition

    You can go to the Expenses and Costs > Order > My Orders page, search by cluster ID, find the order with order type Temporary Upgrade, and click Details to view it.

    Related API

    API

    Description

    TempModifyDBNode

    Temporarily adds nodes to a PolarDB cluster.

    FAQ

    What should I do if I see a "resource unavailable" message when temporarily adding nodes?

    If you select a specific node specification during the upgrade and receive the message Resources unavailable. Try a different region or zone., the selected specification is sold out in the current zone. We recommend:

    1. Select a similar specification: Try a slightly lower specification.

    2. Change the zone: Consider manually switching zones to avoid future resource shortages during configuration changes.

    What should I do if I see a "no specification change" message when temporarily adding nodes?

    You have not added any nodes yet. Follow the on-screen instructions and click Temporarily Add Node or Temporarily Add Columnstore Index Read-Only Node to add a read-only node.

    How can I delete nodes or change compute node configuration for a cluster with temporary nodes?

    You cannot delete nodes or change compute node configuration while temporary nodes are active. First, convert the temporary nodes to permanent nodes, then perform the desired operations.

    What should I do if I see the message "Temporary upgrade expiration time cannot be earlier than: XXX"?

    If performance is still insufficient after temporarily adding nodes and you need to extend the duration, you can perform one additional temporary node addition before the revert time arrives. However, the Restore Point for this second operation cannot be earlier than the first. It must be extended by at least one hour.

    My workload has heavy write pressure. Will adding temporary read-only nodes help?

    No. Read-only nodes handle only read (SELECT) requests. All write operations (INSERT/UPDATE/DELETE) are still processed exclusively by the primary node. If the bottleneck is write performance, upgrade the primary node’s specifications instead of adding read-only nodes.