Data warehouse

In AWS, Amazon Redshift is a data warehouse service that analyzes your data using your existing business intelligence tools. Turbonomic discovers Redshift provisioned clusters through your AWS targets, and represents them as data warehouse entities in the supply chain.

Note:

Turbonomic does not discover Redshift Serverless.

Synopsis

Data warehouse in the supply chain
Synopsis
Provides: Data warehouse services
Consumes: Resources from availability zones
Discovered through: AWS targets

Monitored resources

Turbonomic monitors the following resources:

  • Number of replicas

    Number of replicas is the total number of nodes (leader node and compute nodes) in a Redshift provisioned cluster.

  • Storage amount

    Storage amount is the measurement of storage capacity that is in use.

    Note:

    Turbonomic calculates storage amount. The calculated amount may not match the data returned by Redshift due to an AWS issue. For this reason, you may see a slight difference in the storage amounts shown in Turbonomic and in the AWS console.

  • Connection

    Connection is the measurement of database connections utilized by applications.

Suspend actions for Redshift provisioned clusters

When Turbonomic analysis discovers an idle Redshift provisioned cluster (a cluster with no active connections for a specific period of time), it recommends suspending the cluster as a cost-saving measure and shows the action in Action Center. This action suspends the compute nodes in the cluster, but does not impact data stored on the cluster. In the Action Details page for a specific action, the compute cost is 0 (zero) after suspension. However, you may continue to incur storage cost for certain instance families. For more information about storage cost for instance families, see this AWS page.

To calculate the current compute cost, Turbonomic retrieves the hourly on-demand cost from AWS and then multiplies the result by 730, which represents the number of hours per month that the product uses to estimate monthly costs. Note that for a multi-AZ deployment of a provisioned cluster, the compute cost is double the rate of a single-AZ deployment.

If a currently idle provisioned cluster is not suspended and is subsequently discovered as active, Turbonomic removes the suspend action attached to it.

Suspend actions include the 'Idle time' information that indicates how long a provisioned cluster has been idle. By default, Turbonomic generates suspend actions if a provisioned cluster has been idle for at least 24 hours.

Action Details Idle Time

You can control the suspend actions that Turbonomic generates, based on your preferred idle time value.

For example, if you want Turbonomic to only generate suspend actions for provisioned clusters that have been idle for at least two days, perform these steps:

  1. Create a dynamic group of data warehouses and set the 'Hours Idle' filter to Hours Idle < = 48.

  2. Create a custom data warehouse policy, set the scope to the group that you just created, and then disable suspend actions in that policy.

Action disruptiveness and reversibility

All suspend actions shown in the Action Center view and Action Details page are disruptive but reversible.