Initializing help system before first use

Configuring Worker Resource Usage

Xpress Insight Execution Workers will demand a high proportion of the available resources. To protect the performance of the Insight server, Kubernetes will only schedule Worker pods onto nodes in the cluster that are suitably tainted and labeled. The resource available to these Workers can then be configured.

Tainting Nodes

The Xpress Insight Worker pods use their affinity and toleration properties to define which nodes they will use. Nodes in the cluster can be tainted to allow them to repel a set of pods based on these properties. Tolerations have been applied to allow the Insight Worker pods to schedule onto tainted nodes. Nodes should also be labeled to ensure that Worker pods are not scheduled onto inappropriate nodes, while the taint prevents other pods from being scheduled on these nodes

The FICO® Platform X.x.x Installation Guide section named Prepare the Application Nodes covers the creation of the nodes for the FICO® Platform cluster. Some nodes should be reserved for Insight Execution Workers. For more, see the following topic Configure the number of Workers.

To taint the node, run the following command: 
kubectl taint <node name> dedicated=xpress:NoSchedule
To label it: 
kubectl label nodes <node name> dedicated=xpress
Note: To ensure the all the nodes resources are dedicated to a single worker, workers should be sized to fit one to a node.

Configure the Number of Workers

The resource available to these workers can then be configured from a Linux machine using a bash script that is provided alongside the archetype. The script is named configure-workers.sh. The script allows you to configure the number of workers available and the homogenous resources available across those workers. It also allows you to obtain the current configuration.

To see all the options that are available, open a command prompt and enter the following command: 
 ./configure-workers.sh
-help
Usage
The command is structured as follows: 
 ./configure-workers.sh -d DOMAIN
-s SECRET -c COMPONENT-ID -i INSTANCE-ID -wc VAL -mr VAL -ml VAL
-cr VAL -cl VAL -st VAL -t VAL [-f | --confirm-worker-reduction]
[-p | --print-existing]
Important: During the installation and configuration of FICO® Platform, the installer must configure a clientId and secret for the FICO® Platform "super admin" role — It is this client id and secret that needs to be used to run the configure-workers.sh script.
For example, the following command would provision 1 worker (-wc) with the resources specified in Kubernetes format, with a maximum idle time for the worker of 3600 seconds: 
 ./configure-workers.sh -d
mydmpserver.xpressdev.useast1-fws.ficoanalyticcloud.com -ci
ComponentName -s secretvalue -c componentid -i instanceid -wc 1 -mr
2Gi -ml 3Gi -cr 1m -cl 1 -st 2Gi -t 3600

The domain, client id, and secret can be configured as environment variables to remove the requirement to supply them for each request.

Reducing the Number of Workers

If you need to reduce the number or workers, you must include either -f or -confirm-worker-reduction in the command or it will produce an error.

Parent Topic Editing Execution Worker Settings

© 2001-2022 Fair Isaac Corporation. All rights reserved. This documentation is the property of Fair Isaac Corporation (“FICO”). Receipt or possession of this documentation does not convey rights to disclose, reproduce, make derivative works, use, or allow others to use it except solely for internal evaluation purposes to determine whether to purchase a license to the software described in this documentation, or as otherwise set forth in a written software license agreement between you and FICO (or a FICO affiliate). Use of this documentation and the software described in it must conform strictly to the foregoing permitted uses, and no other use is permitted.