We are implementing an auto allocation of slots to our reservation during peak times. We have commitments for 3 years and we would like this auto allocation to add some slots to increase the max and use auto scaling when needed. I would like to get some idea from the community for the best way to do this – what would be the best conditions and thresholds we should consider deciding when the allocation and de-allocation will happen.
The number of slots you need depends on your performance and workload requirements. You need to understand your workloads to ensure that your BigQuery environment is effectively meeting the needs of your business.
To accomplish this, you can check the number of slots being used by your projects using the following tools:
Administrative Resource Charts: BigQuery administrators can monitor how the organization uses slots and how the job performs every time.
Information Schema: You can query your reservation usage information for detailed information.
Cloud Monitoring and Audit Logs: Leverage this tool to track critical metrics such as slot usage for each reservation and job type.
For a comprehensive understanding about slots and their best practices, refer to the BigQuery documentation. You can also optimize your query to reduce the need for additional slots, improve performance and lower cost.