Updated on 2022-11-02 GMT+08:00

Buying an Instance

Scenario

Kafka instances are physically isolated and exclusively occupied by each tenant. You can customize the computing capabilities and storage space of an instance based on service requirements.

Before You Start

  • Before buying a Kafka instance, ensure that a VPC configured with security groups and subnets is available.
  • (Optional) If you want to access a Kafka instance over a public network, prepare an elastic IP address (EIP) in advance.
  • The broker quantity varies according to the underlying resources, and the underlying resources vary from region to region. Therefore, different instances might have different broker quantities.

Procedure

  1. Log in to the management console.
  2. Click in the upper left corner to select a region.

    Select the same region as your application service.

  3. Click and choose Application > Distributed Message Service (for Kafka) to open the console of DMS for Kafka.
  4. Click Buy Instance in the upper right corner of the page.

    By default, you can create a maximum of 100 Kafka instances for each project. To create more instances, contact customer service to increase your quota.

  5. Specify Billing Mode, Region, Project, and AZ.
  6. Enter an instance name and select an enterprise project.
  7. Configure the following instance parameters:

    Specifications: Select Default or Custom.

    This parameter is available only in the CN East-Shanghai1, CN North-Beijing4, CN South-Guangzhou, AP-Singapore, AP-Jakarta, LA-Mexico City2, LA-Santiago, and AF-Johannesburg regions. In other regions, specifications cannot be customized, and the Default specifications settings are used.

    If you select Default, specify the version, flavor, number of brokers, and broker storage space to be supported by the Kafka instance based on the site requirements.

    1. Version: Kafka v1.1.0, v2.3.0, and v2.7 are supported. v2.7 is recommended. The version cannot be changed once the instance is created.
    2. CPU Architecture: The x86 architecture is supported.
    3. Currently, there are two types of instance flavors depending on the region.
      • In regions other than CN East-Shanghai1, CN North-Beijing4, CN South-Guangzhou, AP-Singapore, AP-Jakarta, LA-Mexico City2, LA-Santiago, and AF-Johannesburg, select a bandwidth based on the service traffic.

        You can view the ECS quantity and flavor, the maximum number of partitions allowed, and number of consumer groups recommended for each bandwidth option.

        The Maximum Partitions parameter indicates the maximum number of partitions that can be created for a Kafka instance. If the total number of partitions of all topics exceeds this threshold, topic creation will fail.

      • In CN East-Shanghai1, CN North-Beijing4, CN South-Guangzhou, AP-Singapore, AP-Jakarta, LA-Mexico City2, LA-Santiago, and AF-Johannesburg regions, select a broker flavor based on the service requirements. For Brokers, specify the broker quantity.

        Maximum number of partitions per broker x Number of brokers = Maximum number of partitions of an instance. If the total number of partitions of all topics exceeds the maximum number of partitions allowed for an instance, topic creation will fail.

    4. Storage Space: Disk type and total disk space for storing the instance data. The disk type cannot be changed once the instance is created.

      The storage space is the total space to be consumed by all replicas. Specify the storage space based on the expected service message size and the number of replicas. For example, if the required disk size to store the data for the retention period is 100 GB, the disk capacity must be at least: 100 GB x Number of replicas + 100 GB (reserved).

      Disks are formatted when an instance is created. As a result, the actual available disk space is 93% to 95% of the total disk space.
      • Flavor kafka.2u4g.cluster: The value range of Storage Space is 600–30,000 GB.
      • Flavor kafka.4u8g.cluster: The value range of Storage Space is 1200–60,000 GB.
      • Flavor kafka.8u16g.cluster: The value range of Storage Space is 2400–900,000 GB.
      • Flavor kafka.12u24g.cluster: The value range of Storage Space is 3600–900,000 GB
      • Flavor kafka.16u32g.cluster: The value range of Storage Space is 4800–900,000 GB
      • 100 MB/s bandwidth: The value range of Storage Space is 600–90,000 GB.
      • 300 MB/s bandwidth: The value range of Storage Space is 1200–90,000 GB.
      • 600 MB/s bandwidth: The value range of Storage Space is 2400–90,000 GB.
      • 1200 MB/s bandwidth: The value range of Storage Space is 4800–90,000 GB.
      • High I/O + 100 MB/s bandwidth: If the average message size is 1 KB, the transactions per second (TPS) can reach 100,000 in high throughput scenarios and 60,000 in synchronous replication scenarios.
      • High I/O + 300 MB/s bandwidth: If the average message size is 1 KB, the TPS can reach 300,000 in high throughput scenarios and 150,000 in synchronous replication scenarios.
      • Ultra-high I/O + 100 MB/s bandwidth: If the average message size is 1 KB, the TPS can reach 100,000 in high throughput scenarios and 80,000 in synchronous replication scenarios.
      • Ultra-high I/O + 300 MB/s bandwidth: If the average message size is 1 KB, the TPS can reach 300,000 in high throughput scenarios and 200,000 in synchronous replication scenarios.
      • Ultra-high I/O + 600 MB/s bandwidth: If the average message size is 1 KB, the TPS can reach 600,000 in high throughput scenarios and 300,000 in synchronous replication scenarios.
      • Ultra-high I/O + 1200 MB/s bandwidth: If the average message size is 1 KB, the TPS can reach 1,200,000 in high throughput scenarios and 400,000 in synchronous replication scenarios.
    5. Capacity Threshold Policy: policy used when the disk usage reaches the threshold. The capacity threshold is 95%.
      • Automatically delete: Messages can be created and retrieved, but 10% of the earliest messages will be deleted to ensure sufficient disk space. This policy is suitable for scenarios where no service interruption can be tolerated. Data may be lost.
      • Stop production: New messages cannot be created, but existing messages can still be retrieved. This policy is suitable for scenarios where no data loss can be tolerated.
    Figure 1 Default specifications

    If you select Custom, the system calculates the number of brokers and broker storage space for different flavors based on your specified peak creation traffic, retrieval traffic, number of replicas per topic, total number of partitions, and size of messages created during the retention period. You can select one of the recommended flavors as required.

    Figure 2 Specification calculation

  8. Configure the instance network parameters.

    • Select a VPC and a subnet.

      A VPC provides an isolated virtual network for your Kafka instances. You can configure and manage the network as required.

      After the Kafka instance is created, its VPC and subnet cannot be changed.

    • Select a security group.

      A security group is a set of rules for accessing a Kafka instance. You can click Manage Security Group to view or create security groups on the network console.

  9. Configure the username and password for logging in to Kafka Manager. The Kafka Manager username cannot be changed once an instance is created.

    Kafka Manager is an open-source tool for managing Kafka clusters. After a Kafka instance is created, you can go to the instance details page to obtain the address for logging in to Kafka Manager. In Kafka Manager, you can view the monitoring statistics and broker information of your Kafka clusters.

  10. Specify the required duration.

    This parameter is displayed only if the billing mode is yearly/monthly.

  11. Click Advanced Settings to configure more parameters.

    1. Configure public access.

      Public access is disabled by default. You can enable or disable it as required.

      After public access is enabled, configure an IPv4 EIP for each broker.

      After enabling Public Access, you can enable or disable Intra-VPC Plaintext Access. If it is enabled, data will be transmitted in plaintext when you connect to the instance through a private network, regardless of whether SASL_SSL is enabled. This setting cannot be changed after the instance is created. Exercise caution. If you want to use a different setting, you must create a new instance.

      • Public access is supported in the following regions: CN North-Beijing4, CN East-Shanghai1, CN South-Guangzhou, CN-Hong Kong, AP-Bangkok, AP-Jakarta, and AP-Singapore.
      • Intra-VPC plaintext access is supported in the CN North-Beijing4 region.
    2. Configure Smart Connect.

      Smart Connect is used for data synchronization between heterogeneous systems. You can configure Smart Connect tasks to dump Kafka data to OBS buckets for backup.

      • You cannot disable Smart Connect once it has been enabled.
      • Smart Connect is supported only in the CN-Hong Kong region.
    3. Configure Kafka SASL_SSL.

      This parameter indicates whether to enable SSL authentication when a client connects to the instance. If you enable Kafka SASL_SSL, data will be encrypted before transmission to enhance security.

      Kafka SASL_SSL is disabled by default. You can enable or disable it as required. This setting cannot be changed after the instance is created. If you want to use a different setting, you must create a new instance.

      If you enable Kafka SASL_SSL, you must also set the username and password for accessing the instance.

    4. Configure Automatic Topic Creation.

      This setting is disabled by default. You can enable or disable it as required.

      If automatic topic creation is enabled, the system automatically creates a topic when a message is created in or retrieved from a topic that does not exist. This topic has the following default settings: 3 partitions, 3 replicas, aging time 72 hours, and synchronous replication and flushing disabled.

      After you change the value of the log.retention.hours, default.replication.factor, or num.partitions parameter, automatically created topics later use the new value. For example, if num.partitions is set to 5, an automatically created topic will have the following settings: 5 partitions, 3 replicas, aging time 72 hours, and synchronous replication and flushing disabled.

    5. Specify Tags.

      Tags are used to identify cloud resources. When you have many cloud resources of the same type, you can use tags to classify them by dimension (for example, use, owner, or environment).

      • If you have predefined tags, select a predefined pair of tag key and value. Click View predefined tags. On the Tag Management Service (TMS) console, view predefined tags or create tags.
      • You can also create new tags by specifying Tag key and Tag value.

      Up to 20 tags can be added to each Kafka instance. For details about the requirements on tags, see Managing Instance Tags.

      Some regions do not support tags. For details, see the information displayed on the console.

    6. Enter a description of the instance.

  12. Click Buy.
  13. Confirm the instance information, and read and agree to the HUAWEI CLOUD Customer Agreement. If you have selected the yearly/monthly billing mode, click Pay Now and make the payment as prompted. If you have selected the pay-per-use mode, click Submit.
  14. Return to the instance list and check whether the Kafka instance has been created.

    It takes 3 to 15 minutes to create an instance. During this period, the instance status is Creating.

    • If the instance is created successfully, its status changes to Running.
    • If the instance fails to be created, view Instance Creation Failures. Delete the instance by referring to Deleting an Instance and create another instance. If the instance creation fails again, contact customer service.

      Instances that fail to be created do not occupy other resources.