Changing Query Engine instance type capacity
You can optimize cluster performance and cost by adjusting the instance type in response to instances that are underutilized or overutilized. This flexibility enables more efficient resource utilization and cost management, tailored to dynamic workload requirements.
Changing the instance type capacity for Query Engine allows:
Scale Up: You can increase the instance type size beyond the initially deployed configuration to accommodate higher processing requirements. For example, if a cluster was originally deployed with a maximum capacity of 16 GB RAM, it can now be changed to 128 GB RAM to deliver enhanced processing performance.
Scale Down: You can decrease the instance type size to optimize costs during periods of lower demand. For example, if a cluster was deployed initially with a maximum capacity of 128 GB RAM, it can now be changed to 16 GB RAM for the use of even smaller instances, thereby improving cost efficiency.
Note
Changing the instance type is only allowed within the same category. You cannot switch to an instance type that belongs to a different category than the one currently in use.
A restart of Kyvos services is required.
You may need to adjust the schedules based on the changes made to the maximum capacity.
Ensure that changes to the Query Engine instance type via the Kyvos Manager (KM) are not applied in real time. Specifically, modifying the Query Engine instance type does not immediately affect the currently running Query Engine instances. The new instance type will only take effect after a restart of the Kyvos services. Even if a new schedule is triggered post-update, the change will not apply until the services are restarted.
To change the Query Engine instance capacity, perform the following steps.
On the navigation pane, click Kyvos and Ecosystem> Nodes and Services > Overview.
On the Overview page, you will see the number of nodes for Kyvos Core Services and the number of service instances (BI Server, Query Engine, and Web Portal) configured for the Kyvos cluster.
In the Query Engine tile, click Change Instance Capacity. The Change Query Engine Instance Capacity dialog is displayed.
Select the Instance Type from the list. It will correspond to the required capacity of a Query Engine.
Stop all running Query Engines and BI Server. To stop them, navigate to the Cluster Scaling page in the Kyvos Portal. Click the Change Cluster State link. On the Change Cluster State dialog, click the Cluster state as offline. This will stop the Query Engine and BI Server.
Note
After stopping the Query Engine and BI Server, the cluster will return to an active state after approximately one hour.
Enter your Kyvos Manager password.
Click Change.
Important
It is recommended that your cluster scaling schedule in the Kyvos portal be updated to reflect the new Query Engine capacity.