Pre-Requisites

Software Pre-requisites

• Minimum Requirements

  • IBM FileNet Software (Images)
  • Kubectl
  • AWS CLI
  • IAM
    • AWS EKS Security Group should ALLOW communication on assigned NodePorts

• Hardware

  • EKS
    • m5.xlarge and region as us-east-1. (this has 4 vcpu and 16 gigs ram)
    • Default storage class defined
  • Jump Server/Bastion Host for staging requirements

Cluster Sizing

Small (targets development)

Component	CPU Request (m)	CPU Limit (m)	Memory Request (Mi)	Memory Limit (Mi)	Number of replicas
CPE	1000	1000	3072	3072	1
CSS	1000	1000	4096	4096	1
CSGraphQL	500	1000	1536	1536	1
Navigator	1000	1000	3072	3072	1
External Share	500	1000	1536	1536	1
Task Manager	500	1000	1536	1536	1
CMIS	500	1000	1536	1536	1

Medium (targets production with high-availability)

Component	CPU Request (m)	CPU Limit (m)	Memory Request (Mi)	Memory Limit (Mi)	Number of replicas
CPE	1500	2000	3072	3072	2
CSS	1000	2000	8192	8192	2
CSGraphQL	500	2000	3072	3072	2
Navigator	2000	3000	4096	4096	2
External Share	500	1000	1536	1536	2
Task Manager	500	1000	1536	1536	2
CMIS	500	1000	1536	1536	2

Large (targets production with high-availability)

Component	CPU Request (m)	CPU Limit (m)	Memory Request (Mi)	Memory Limit (Mi)	Number of replicas
CPE	3000	4000	8192	8192	2
CSS	2000	2000	8192	8192	2
CSGraphQL	1000	2000	3072	3072	6
Navigator	2000	4000	6144	6144	6
External Share	500	1000	1536	1536	2
Task Manager	500	1000	1536	1536	2
CMIS	500	1000	1536	1536	2

We are going with MEDIUM, so technically we only need 12 vcpus and 24 gigs RAM. We are going to provision a 6 node cluster with the m5.xlarge sizing as this will give us 24 vcpu and 96 gigs RAM, which is way more than we need anyway. Man, if there was only a sizing in AWS that was 4 VCPU/8 GIGS, that would be ideal as we're going to have twice as many vpcus and 4 times as much RAM as we need.