2.9.2 Provisioning

To determine the appropriate HeatWave Cluster size for a workload, you can estimate the required cluster size. Cluster size estimates are generated by the HeatWave Auto Provisioning feature, which uses machine learning models to predict the number of required nodes based on node shape and data sampling. For instructions:

• For MySQL HeatWave on OCI, refer to Generating a Node Count Estimate, in the MySQL HeatWave Service User Guide.

• For MySQL HeatWave on AWS, see Estimating Cluster Size with MySQL Autopilot, in the MySQL HeatWave on AWS Service Guide.

• For HeatWave in Oracle Database Service for Azure (ODSA), see Provisioning HeatWave Nodes.

Perform a cluster size estimate:

• When adding a HeatWave Cluster to a DB System, to determine the number of nodes required for the data you intend to load.

• Periodically, to ensure that you have an appropriate number of HeatWave nodes for your data. Over time, data size may increase or decrease, so it is important to monitor the size of your data by performing cluster size estimates.

• When encountering out-of-memory errors while running queries. In this case, the HeatWave Cluster may not have sufficient memory capacity.

• When the data growth rate is high.

• When the transaction rate (the rate of updates and inserts) is high.