Identifying the infrastructure requirements

All the Cloud Pak containers are based on Red Hat Universal Base Images (UBI), and are Red Hat and IBM certified. To use the Cloud Pak images, it is important that you understand what you must do before you install the Cloud Pak operator.

For each stage in your operations (a minimum of three stages is expected development, preproduction, and production), you must allocate a cluster of nodes before you install the Cloud Pak. Development, preproduction, and production are stages that are best run on different compute nodes.

Note: Use the shared_configuration.sc_deployment_license parameter to define the purpose of the "enterprise" deployment type (shared_configuration.sc_deployment_type). Valid values are production and non-production.

The Detailed system requirements page provides a cluster requirements guideline for IBM Cloud Pak® for Automation.

The minimum cluster configuration and physical resources that are needed to run the Cloud Pak includes (but is not limited to):

  • Hardware architecture: Intel (amd64 or x86_64 the 64-bit edition for Linux® x86) on all platforms.
  • Node counts: Dual compute nodes for non-production and production clusters. A minimum of three nodes is needed for medium and large production environments and larger test environments. Any cluster configuration needs to adapt to the size of the projects and the work load that is expected.
 For 20.0.2  A cluster where you want to install all of the capabilities needs as a minimum:
  • Master(3 nodes): 4 vCPU and 8 Gi memory on each node.
  • Worker(3 nodes): 16 vCPU and 32 Gi memory on each node.
 New in 20.0.3  A cluster where you want to install all of the capabilities needs as a minimum:
  • Master (3 nodes): 4vCPU and 8 Gi memory on each node.
  • Worker (5 nodes): 16vCPU and 32 Gi memory on each node.
Note: Business Automation Insights can coexist with another Cloud Pak in different namespaces on a cluster, provided the cluster has sufficient capacity and that both Cloud Paks use the same version of Common Services.

The following tables provide minimum resources for each capability. If you plan to install Business Automation Insights, you must install IBM Event Streams beforehand. For more information about the minimum requirements of Event Streams, see https://ibm.github.io/event-streams/. Event Streams needs an instance of Common Services on the cluster. For more information about the minimum requirements of Common Services, see Hardware requirements and recommendations for common services.

  • Table 1 Cloud Pak for Automation operator minimum requirements
  • Table 2 Automation Decision Services minimum requirements
  • Table 3  New in 20.0.3  Automation Document Processing minimum requirements
  • Table 4  New in 20.0.3  Automation Workstream Services minimum requirements
  • Table 5 Business Automation Application minimum requirements
  • Table 6 Business Automation Insights minimum requirements
  • Table 7 Business Automation Navigator minimum requirements
  • Table 8 Business Automation Studio minimum requirements
  • Table 9  New in 20.0.3  Business Automation Workflow minimum requirements with or without Automation Workstream Services
  • Table 10 FileNet® Content Manager minimum requirements
  • Table 11 Operational Decision Manager minimum requirements
  • Table 12 User Management Service minimum requirements
Removed or replaced capabilities:
  • Table 13  For 20.0.2  Automation Content Analyzer minimum requirements
  • Table 14  Removed in 20.0.3  Automation Digital Worker minimum requirements
  • Table 15  For 20.0.2  Automation Workstream Services and Business Automation Workflow minimum requirements
  • Table 16  For 20.0.2  Enterprise Records minimum requirements
Table 1. Cloud Pak for Automation operator minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
ibm-cp4a-operator 500 1000 256 1024 1 No
Table 2. Automation Decision Services minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
ads-runtime 1000 2000 512 2048

 For 20.0.2  2

 New in 20.0.3  from 2 to 5 with horizontal pod autoscaling enabled

 Yes
 New in 20.0.3  ads-embedded-runtime 1000 2000 512 2048 2 Yes
ads-download 100  For 20.0.2  100

 New in 20.0.3  500

  For 20.0.2  128

 New in 20.0.3  200

  For 20.0.2  128

 New in 20.0.3  200

  For 20.0.2  1

 New in 20.0.3  2

 No
ads-front 500 2000 512 2048 2 No
ads-gitservice 500 2000 512 2048 2 No
ads-parsing 500 2000 512 2048 2 No
ads-restapi 500 2000 512 2048 2 No
ads-run 500 2000 512 2048 2 No
 New in 20.0.3  ads-build 500 2000 512 2048 2 No
 New in 20.0.3  ads-credentials 500 2000 512 2048 2 No
mongo 500 1000 256 1024 2 No
Note: Automation Decision Services also creates some jobs that request 200m CPU and 256Mi Memory. The following jobs are created once at the beginning of the installation and do not last long:
  • ads-ltpa-creation
  • ads-runtime-bai-registration
  • ads-runtime-ums-registration
  • ads-ums-registration

The ads-rr-integration job is started every 15 minutes, and is also short-lived.

Table 3.  New in 20.0.3  Automation Document Processing minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of Replicas Pods are licensed for production/non-production
OCR Extraction 200 1000 2048 4096 2 Yes
Classify Process 200 1000 100 960 2 Yes
Processing Extraction 500 1000 1024 2048 2 Yes
Natural Language Extractor 200 1000 100 1440 2 Yes
Common Git Gateway Service (git-service) 500 1000 512 3072 2 No
Content Designer Repo API (CDRA) 500 1000 512 1024 2 No
Content Designer UI & REST (CDS) 500 1000 512 1024 2 No
Content Project Deployment Service (CPDS) 500 1000 512 1024 2 No
Mongo database (mongodb) 100 500 256 512 1 No
Viewer service (viewone) 500 1000 1024 4096 2 No
Callerapi 200 600 1024 3072 2 No
PostProcessing 200 600 100 480 2 No
PDFProcess 200 600 100 960 2 No
Utf8Process 200 1000 100 960 2 No
Setup 200 600 100 700 2 No
Deep Learning 1000 2000 3072 6144 2 No
UpdateFileDetail 200 600 100 480 2 No
Backend 200 600 100 640 2 No
Redis 100 250 100 640 3 No
RabbitMQ 100 500 100 640 3 No
Important: Document Processing - For Deep Learning, it is highly recommended to use worker nodes with NVIDIA GPU for better and faster results. NVIDIA is the only supported GPU for Deep Learning in the Document Processing pattern. Use these installation instructions to install the NVIDIA GPU Operator.

The GPU worker nodes must have a unique label, like ibm-cloud.kubernetes.io/gpu-enabled:true. You add this label value to the deployment script or your CR YAML when you configure the YAML for deployment. In order to achieve HA, you need a minimum of 2 GPU so that 2 replicas of Deep Learning pods can be started. You can change the replica to 1 if you only have 1 GPU on the node.

Table 4.  New in 20.0.3  Automation Workstream Services minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
Workflow Server 500 2000 1024 2560 2 Yes
Java™ Message Service 200 1000 512 2048 1 No
Process Federation Service 500 2000 512 4096 1 No
Process Federation Service-dbareg 200 500 128 256 1 No
Elasticsearch Service 100 1000 2048 2048 1 No
Table 5. Business Automation Application minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
App Engine 300 2000 512 2048 1

 New in 20.0.3  2

 Yes/No
Resource Registry 100 500 128 512 3 No
Table 6. Business Automation Insights minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
Flink task managers 1000 1000 1280 1536 Default parallelism

 For 20.0.2  1

 New in 20.0.3  2

 Yes/No
Business Performance Center  For 20.0.2  Not specified

 New in 20.0.3  100

 4000 512 2000 1 Yes/No
Flink job manager 100 - 256 1280 1 No
Flink Zookeeper 100 200 640 768  For 20.0.2  1

 New in 20.0.3  2

 No
Administration REST API 3 - 50 120  For 20.0.2  1

 New in 20.0.3  2

 No
 New in 20.0.3  Management REST API 3 - 50 120  New in 20.0.3  2 No
Elasticsearch client 100 1000 1000 2000  For 20.0.2  1

 New in 20.0.3  2

 No
Elasticsearch master 100 1000 1000 2000  For 20.0.2  1

 New in 20.0.3  2

 No
Elasticsearch data 100 1000 1000 2000  For 20.0.2  1

 New in 20.0.3  2

 No
Kibana 100 1000 1000 2000 1 No
Note: Business Automation Insights also creates a Setup job that requests 200m CPU and 50Mi Memory. A pod is created at the beginning of the installation and does not last long.
Table 7. Business Automation Navigator minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
Navigator 500 1000 512 1536 2 No
Table 8. Business Automation Studio minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
App Engine playback 300 2000 512 2048 1

 New in 20.0.3  2

 Yes/No
BAStudio 1000 4000 1024 3072 1 No
BAStudio JMS 200 1000 256 1024 1 No
Resource Registry 100 500 128 512 3 No
Table 9.  New in 20.0.3  Business Automation Workflow minimum requirements with or without Automation Workstream Services
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
Workflow Server 500 2000 1024 2560 2 Yes
Workflow Authoring 500 4000 1024 3072 1 No
Java Message Service 200 1000 512 2048 1 No
Process Federation Service 500 2000 512 4096 1 No
Process Federation Service-dbareg 200 500 128 256 1 No
Elasticsearch Service 100 1000 2048 2048 1 No
Intelligent Task Prioritization 500 2000 1024 2048 1 No
Workforce Insights 500 2000 1024 2048 1 No
Table 10. FileNet Content Manager minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
CPE 500 1000 512 3072 2 Yes
CSS 500 1000 512 4096 2 Yes
 New in 20.0.3  Enterprise Records (ER) 500 1000 512 1536 1 Yes
 New in 20.0.3  Content Collector for SAP (CC4SAP) 500 1000 512 1536 2 Yes
CMIS 500 1000 256 1536 2 No
GraphQL 500 1000 512 1536 2 No
External Share 500 1000 512 1536 2 No
Task Manager 500 1000 512 1536 2 No
Table 11. Operational Decision Manager minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
Decision Center 500 2000 1500 4096 1

 New in 20.0.3  2

 Yes
Decision Runner 500 1000 512 4096 1

 New in 20.0.3  2

 Yes
Decision Server Runtime 500 2000 512 4096 1

 New in 20.0.3  2

 Yes
Decision Server Console 500 2000 512 1024 1

 New in 20.0.3  2

 No
Note: Operational Decision Manager also creates an odm-oidc-job-registration job that requests 200m CPU and 256Mi Memory. The pod is created at the beginning of the installation and does not last long.
Table 12. User Management Service minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
UMS 200 500 256 512 2 No

To achieve high availability, you must adapt the cluster configuration and physical resources. You can set up a Db2® High Availability Disaster Recovery (HADR) database. For more information, see Setting up Db2 High-Availability. For high availability and fault tolerance to be effective, set the number of replicas you need for the respective configuration parameters in your custom resource file and let the operator manage the scaling.

Removed or replaced capabilities

Table 13.  For 20.0.2  Automation Content Analyzer minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
OCR Extraction 200 1000 100 1440 2 Yes
Classify Process 200 1000 100 960 2 Yes
Processing Extraction 200 1000 100 1440 2 Yes
Callerapi 200 600 100 1024 2 Yes
PostProcessing 200 600 100 480 2 Yes
PDFProcess 200 600 100 960 2 Yes
Utf8Process 200 1000 100 960 2 Yes
Setup 200 600 100 480 2 Yes
UpdateFileDetail 200 600 100 480 2 Yes
Backend 200 600 100 640 2 Yes
Frontend 200 600 100 480 2 Yes
Redis 200 500 200 1280 3 Yes
Rabbit MQ 100 500 100 640 3 Yes
Table 14.  Removed in 20.0.3  Automation Digital Worker minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
Digital Worker Designer 200 1000 256 2000 1 Yes
Digital Worker Tasks Runtime 500 1000 256 2000 2 Yes
Digital Worker Management Server 500 1000 512 2000 1 Yes
Digital Worker Set up Job 200 200 256 2000 1 and only 1 Yes
MongoDB 500 1000 256 2000 2 Yes
NPM registry 500 1000 256 2000 1 Yes
Table 15.  For 20.0.2  Automation Workstream Services and Business Automation Workflow minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
Workflow Server 500 2000 1048 2096 2 Yes
Java Message Service 200 1000 512 2048 1 No
Process Federation Service 500 2000 512 4000 1 No
Process Federation Service-dbareg 500 2000 512 4096 1 No
Elasticsearch Service 100 1000 1024 2048 1 No
Table 16.  For 20.0.2  Enterprise Records minimum requirements
Component CPU Request (m) CPU Limit (m) Memory Request (Mi) Memory Limit (Mi) Number of replicas Pods are licensed for production/non-production
IER 500 1000 512 3072 1 Yes