This topic describes how to scale out or scale in a Realtime Compute for Apache Flink cluster or project.

Change resource configurations of a cluster

Note
  • Fees after the resource configuration change:
    • Scale-out (Master Scale-Up or Slave Scale-Out): You must pay the relevant upgrade fees.
    • Scale-in (Master Scale-Down or Slave Scale-In): The refund from the scale-in is returned to your account.
  • After you perform Master Scale-Up or Slave Scale-Out, you must add required IP addresses to a whitelist of the database. For more information, see Configure a whitelist for accessing storage resources.

You can modify the configuration of Master Specifications or Slave Quantity to change the configuration of the cluster.

  1. Go to the Project Management page.
    1. Log on to the Realtime Compute development platform.
    2. Move the pointer over the username in the upper-right corner.
    3. Click Project Management.
  2. In the left-side navigation pane, choose Cluster Management > Clusters.
  3. On the Clusters page, find your cluster, click More in the Actions column, and select Scale Out or Scale In.
  4. Change resource configurations.
    The following example describes two scale-out methods:
    • Master Scale-Up (use higher specifications)
      1. On the Update page, select Master Scale-Up for Upgrade Method.
      2. Specify Master Specifications.
      Note
      • You can change the master specifications but cannot change the number of master nodes.
      • If you use a high availability cluster configured with three master nodes, the system upgrades all three master nodes at the same time when you select Master Scale-Up.
    • Slave Scale-Out (increase the number of slave nodes)
      1. On the Update page, select Slave Scale-Out for Upgrade Method.
      2. Increase the value of Slave Quantity.
      Note
      • If the number of CUs is insufficient, we recommend that you select Slave Scale-Out and increase the value of Slave Quantity.
      • If you are unable to complete payment after you specify Slave Quantity and read and select Realtime Compute Exclusive Mode (Subscription) Terms of Service, add a vSwitch and enter the ID of the new vSwitch in vSwitchId. After verification, you can complete the payment for the scale-out. For more information about how to add a vSwitch, see Work with vSwitches.
      • The slave nodes that you want to add must be located in the same zone as the cluster.
      • You cannot change the slave specifications for an existing Realtime Compute for Apache Flink cluster in exclusive mode. You can only increase the value of Slave Quantity. If you want to change the slave specifications, you must purchase a new cluster.
  5. Read and select Realtime Compute Exclusive Mode (Subscription) Terms of Service.
  6. Click Pay.

Change resource configurations of a project

Note
  • If the computing capability of existing resources does not meet your business requirements, you can scale out the project to improve the computing capability of the system.
  • If the computing capability of existing computing resources is excessive to meet your business requirements, you can downgrade the resource configuration to reduce costs.
  1. Go to the Project Management page.
    1. Log on to the Realtime Compute development platform.
    2. Move the pointer over the username in the upper-right corner.
    3. Click Project Management.
  2. In the left-side navigation pane, choose Project Management > Projects.
  3. Change resource configurations:
    1. On the Projects page, find your project and click Scale In/Out in the Actions column.
    2. In the Scale In/Out dialog box, specify Specified CUs.
    3. Click OK.