Overview
If the current instance specifications cannot meet your business needs, you can upgrade them in the console.
Note:
You can upgrade instance specifications via the console, including increasing peak bandwidth, disk capacity, and the number of partitions. The upgrade process is a seamless transition, ensuring that your service remains uninterrupted.
New ports may be added during the upgrade. You can view them by clicking View All IPs and Ports in the Access Mode section on the instance details page later. When configuring the security group, you need to open all these ports.
Prerequisites
Before upgrading, please check the following items:
1. Check whether the instance has unavailable public network routes, supporting networks, and VPC networks. For more information, see Adding Routing Policy. 2. Check whether the instance has unsynced replicas. For more information, see Viewing Topic. 3. Check whether the instance has unfinished tasks (e.g., data migration), abnormally created topics, or abnormally deleted topic data.
Note:
If there are unfinished tasks, it is recommended to wait until all are completed before proceeding with the upgrade. If any task encounters an exception, submit a ticket for assistance. Directions
2. On the Instance List page, select More > Upgrade from the operation column to enter the Instance Upgrade Page.
3. On the Instance Upgrade Page, select the Target Upgrade Specification.
Product specification: Select a model based on the peak bandwidth and disk capacity.
Instance price: If you choose to upgrade, you will need to pay the daily difference in price. For more details, see Monthly Subscription and Product Configuration Upgrade Instructions.
Rebalance Time:
If the backend detects that data migration is required during the upgrade, you can choose to upgrade immediately or at a defined time (It is recommended to upgrade at nighttime to minimize the business impact.). The estimated duration is calculated by backend APIs.
If the backend detects that data migration is not required during the upgrade, This Upgrade will not Trigger a Rebalance will be displayed.
Upgrade mode: If the backend detects that instance migration is required during the upgrade, you can select an upgrade mode as needed. Otherwise, you do not need to select an upgrade mode.
Stable mode: In Stable Mode, CKafka will limit the speed of data migration during the upgrade to preserve the instance's bandwidth attributes. This option is suitable for scenarios where you do not want to affect business operations.
High-Speed mode: In high-speed mode, CKafka will not limit the speed of data migration during the upgrade, which may affect the bandwidth for production and consumption of instances. This option is suitable for low business peak hours or scenarios where service suspension is acceptable.
4. Click Submit to complete the instance upgrade. You can view the real-time progress of the upgrade in the Status Column.
5. If a scheduled upgrade was set, you can modify the scheduled time in the Status Bar.
Possible Causes of Upgrade Failure
1. One possible reason is that the disk resources in the current AZ cannot meet the requirements of this upgrade. It is recommended to contact the Tencent Cloud customer service to check whether there are sufficient resources.
2. If the high-speed mode is selected during the instance upgrade, and there are bandwidth-intensive production tasks in the cluster, the data migration delay will increase. You can view monitoring data to see whether there are excessive peaks of production and consumption traffic during the upgrade. For more information, see Viewing Monitoring Data. 3. The upgrade process takes too long because the maximum message size configured for the broker being migrated is 8 MB, while the target broker is set to 1 MB. This causes the target broker to be unable to receive oversized messages, leading to data migration failure and extending the upgrade duration. You can submit a ticket for assistance. 4. During the upgrade or migration of a cluster, an exception may occur while updating the broker IP, causing the new cluster to fail to pull the broker IP data. If you observe no monitoring data for an extended period, please submit a ticket for assistance. For more information, see Viewing Monitoring Data. Non-operational Functions During Upgrade
During cluster reconfiguration, it is recommended that you do not operate Topic management-related functions, such as creating Topics or editing Topic attributes. The list of related functions is as follows.
Function Module | Specific Functions | Description |
Topic management list | Create a Topic | Create a Topic under the cluster. |
| Edit a Topic | Edit Topic attributes under the cluster, such as changing the number of partitions or replicas. |
| Delete a Topic | Delete a Topic under the cluster. |
| Throttling | Configure throttling policies for Topics. |
| Manual partition balancing | Manually perform partition balancing to resolve uneven partition allocation issues. |
Smart ops (Professional Edition exclusive feature) | Disk water level processing | Change dynamic message retention policy, or automatic disk expansion, switch status. |
| Dynamic partition processing | Perform manual partition balancing diversion, or configure automated partition balancing policy. |
| Bandwidth elastic configuration | Change the switch status of elastic bandwidth |
