Automatically Pausing and Resuming Serverless Instances

Scenarios

When using TaurusDB serverless instances, you can flexibly adapt to changes in workload demand through on-demand scaling and pay-per-use billing. However, during development and testing or off-peak hours at night, instances may continue running while remaining idle for extended periods, incurring unnecessary compute costs. TaurusDB serverless instances support automatic pause and resume. Once the feature is enabled, if the compute capacity of every node in an instance does not exceed 2 TCUs and there are no active user connections for the configured idle duration, the instance automatically scales its compute capacity down to zero and compute billing stops. When a new user connection request is detected, the instance automatically wakes up and billing resumes. No manual intervention is required throughout the process.

Billing

When compute capacity scales down to zero, compute nodes stop incurring charges. Other associated resources, such as EIPs and backups, may continue to incur charges. For details, see Billing.

Constraints

This feature is available only for serverless instances.

Procedure

1. Log in to the TaurusDB console.
2. On the Instances page, click the instance name to go to the Basic Information page.
3. In the Serverless Information area, click Configure under Auto Pause/Resume.
   Figure 1 Modifying serverless information
   

4. In the displayed dialog box, click to set the detection duration.
   Figure 2 Configuring auto pause/resume
   

5. Click OK to enable auto pause/resume.
6. When the compute capacity of every node in the instance does not exceed 2 TCUs and there are no active user connections for the configured detection duration, the compute capacity automatically scales down to zero. You can confirm that the compute capacity has dropped to 0 by checking monitoring metrics (CPU count).
   Figure 3 Checking monitoring metrics