Monitor PCU groups in Astra

Astra provides insight into your PCU usage, including overall group usage and individual database activity.

You can monitor your PCU usage to ensure there are sufficient resources to support your databases and to manage costs by avoiding over-provisioning resources.

Prerequisites

Understand resource consumption in PCU groups

Databases in a PCU group use the group’s minimum capacity first.

If database usage scales beyond the minimum capacity, Astra can provision additional PCUs, billed at the HCU rate up to the group’s maximum capacity. As usage scales down, Astra deprovisions the excess PCUs accordingly.

If there are not enough resources available, database operations can experience higher latencies or timeouts.

If your PCU group contains Serverless (Non-Vector) databases, and database demand bursts the group’s maximum capacity, then DataStax may increase the maximum capacity on your behalf if your databases are at risk of severe performance degradation.

DataStax will notify you if this occurs, and then you can choose to maintain the increased maximum capacity or revert the maximum capacity to its original value.

Burst usage is the same as regular maximum capacity usage: it only incurs costs while in use. Once demand decreases below the burst point, Astra automatically deprovisions this additional capacity. However, the PCU group’s maximum capacity remains at the increased value unless you decrease it by editing the PCU group.

Burst capacity doesn’t apply to PCU groups with Serverless (Vector) databases.

View PCU group usage

To view usage for an entire PCU group, do the following:

In the Astra Portal header, click Settings.
In the Settings navigation menu, click the name of the active organization, and then select the organization where you want to view PCU group usage.

If the organization belongs to an enterprise, select the enterprise, and then select the organization in the Organizations list.
In the Settings navigation menu, click Provisioned capacity.
Find the group you want to inspect, click More, and then select View details to inspect the group’s configuration details, scoped databases, and usage metrics.

PCU group details

This section contains a summary of the PCU group’s configuration, such as the following:

Name and description
Cloud provider and region
Workload type
Cache storage
PCU allocation

Scoped databases

This section lists the Name, Datacenter ID, and Status for each database assigned to this PCU group.

PCU group metrics describe activity for the entire PCU group. To determine which databases are the most active and resource-intensive in a PCU group, you must inspect individual database usage. For more information, see View and export metrics.

Overall PCU utilization

This section provides graphs of key PCU utilization metrics, including the following:

Performance: PCU group performance over time, represented by CPU usage, latency, and cache utilization.
- CPU: The amount of CPU that the PCU group is using out of the group’s total amount of available CPU.
- Latency: The time required for reads and writes against each database in the group.
- Cache utilization: The amount of cache capacity the PCU group is currently using out of the group’s total cache capacity.
Capacity: Scaling activity over time.

Reserved and minimum capacity are always in use. The type (RCU or HCU) depends on the PCU group’s Workload type and capacity allocation.

Scaling events indicate periods in which the PCU group scaled into the maximum capacity or triggered a burst capacity event.

For more information, see Understand resource consumption in PCU groups.

To view more granular point-in-time metrics, hover over the lines and bars on the graphs.

To change the reporting period, use the Time Frame Section menu. The default time range is the last 7 days.

Respond to usage concerns

After inspecting the PCU group and database usage, you might decide that you need to modify your PCU groups. For example:

To balance resource consumption and availability, you can edit your PCU groups to reassign databases or adjust allocated capacity.
For resource-intensive or critical production workloads, DataStax recommends creating PCU groups dedicated to individual databases.
To reduce costs, you can park flexible capacity PCU groups when you don’t need those resources to be active.