We have one VA cluster with two servers VA1 and VA2. Last week we integrated 6 more AD Domains. Now, intotal we have 40 applications, in which we have 8 AD domains running. And we are planning to add 10 more applications.
Both VAs has 2 processors, 16GB Memory and 128GB storage
Now our authoritative source Oracle Cloud is taking around 5 hours to complete the aggregation and sometimes it will error as time out.
We are assuming it is because of one VA cluster with 2 servers. We decided to add 2 more servers. But i have question.
Option1
Add two servers to the existing VA cluster (in total we have 4 servers)
Option2
Create a second VA cluster with the 2 servers ( we will have two VA clusters with 2 servers each) — distribute all applications between 2 Clusters.
Should I opt for option 1 or 2?
And questino2: If there are x number of servers in a single cluster will SailPoint load balance the number of tasks between these x number of servers?
When there are multiple VAs in a cluster, each will try to contact the tenant to check for work in the queue. So, if there are two aggregations that are scheduled, VA1 could start the first aggregation and V2 could start the second aggregation. Adding servers could help by allowing multiple processes to run simultaneously.
Before deciding option 1 or option 2, how frequently are you aggregating your sources? I am wondering if they are getting backed up. For example, I might consider aggregating the HR source more frequently so that hires and terminations could be processed more timely but maybe doing an application that I just use for certification less frequently.
As far as dividing them into two clusters, four servers can aggregate four sources. You can divide them into two clusters and evenly divide the sources or you could also think about Cluster 1 being the priority sources - HR, , … and then putting more on Cluster 2 that just need to get done at some point.
Thank you for your detailed response, Alicia. It has provided valuable insight into the topic.
We currently manage three applications sourced from HR. Initially, we aggregate data from these applications 4 times a day. However, one of the application, which uses a direct connector (oracle Fusion HCM- we’re still on the old connector but plan to upgrade with in the next two months), was taking about 4-5 hours to complete the aggregation. As a result, we reduced the frequency to twice a day.
Additionally, we run aggregations for our CIEM sources, JDBC, SAP and others once a day, scheduled them sequentially throughout the day.
I’m still uncertain whether to proceed with option 1 (adding 2 more servers to the existing cluster) or option 2 ( create a new cluster with 2 new servers and share the sources among them). Could you please advice on the best course of action?