Scaling Product Master instances

The Product Master service supports four standard installation sizes, which you configure at the time of installation by using the custom resource. You can scale Product Master service based on usage to support high-availability, or increase processing capacity.

The following table describes the t-shirt sizing for the Product Master service.
Minimum requirements
14 vCPU | 48 GB RAM | 200 GB Storage
Standard
  vCPU RAM (GB) Storage(GB) Description
Small 20 56 250
  • Designed to support up to 300,000 items with 30 concurrent users.
  • Minimum CPU and RAM request values are 14 vCPU and 44 GB.
Medium 34 88 500
  • Designed to support up to 1,000,000 items with 30 concurrent users.
  • Minimum CPU and RAM request values are 24 vCPU and 68 GB.
Large 64 128 1 TB
  • Designed to support up to 5,000,000 items with 85 concurrent users.
  • Minimum CPU and RAM request values are 44 vCPU and 100 GB.
XLarge

96

200

 5 TB
  • Designed to support up to 30,000,000 items with 150 concurrent users.
  • Minimum CPU and RAM request values are 68 vCPU and 156 GB.
Note: Estimates can vary depending on change to data models such as number of attributes, matching algorithm, or data quality (percentage of duplicates, and so on).

If the fixed t-shirt sizing is not appropriate or does not provide adequate capacity (for example, you need something bigger than the large configuration), you can use custom sizing.

To enable HPA after Product Master deployment,
  • Set the value of autoScaleConfig property as true in the CR file as follows.
    cat <<EOF |oc apply -f
apiVersion: productmaster.cpd.ibm.com/v1
kind: ProductMaster
metadata:
  name: productmaster-cr
  labels:
    app.kubernetes.io/instance: ibm-cpd-productmaster-operator
    app.kubernetes.io/managed-by: ibm-cpd-productmaster-operator
    app.kubernetes.io/name: ibm-cpd-productmaster-operator
spec:
  version:  "3.2.0"
  scaleConfig: small
  license:
    accept: false
  shutdown: false
  #To enable HPA, autoScaleConfig should be set to true
  autoScaleConfig: <true/false>
  fileStorageClass: <file storage name>
  blockStorageClass: <block storage name>
EOF
    The following table shows the values (HPA name, Average utilization metric, Minimum replica, and Maximum replica) for each service to be specified for the respective t-shirt sizes.
    T-shirt sizes Small Medium Large XLarge
    HPA NAME Average utilization metric Minimum replica Maximum replica Minimum replica Maximum replica Minimum replica Maximum replica Minimum replica Maximum replica
    hpa-admin 50 2 5 2 6 2 8 2 10
    hpa-personaui 50 2 2 2 3 2 4 2 5
    hpa-ftsind 50 2 2 2 3 2 4 2 5
    hpa-ftspim 50 2 2 2 3 2 4 2 5
    hpa-wfl 50 2 5 2 6 2 8 2 10
    hpa-sch 50 2 5 2 6 2 8 2 10
    hpa-gds 50 2 5 2 6 2 8 2 10
    hpa-ml 50 2 5 1 1 1 1 1 1
    hpa-restapi 50 2 5 2 6 2 8 2 10