Hi! I am working in a project, where there are 81 JDBC connectors. Production cluster has 2 VA, each with 2 processors and 8GB RAM. Each connector has about 900 entitlement and 12k accounts. Each connector aggregation takes about 30 minutes.
I notice that when several aggregation overlaps, other subsequent connectors that starts their aggregation become stuck, and ends in timeout.
This cluster is exclusive for this 81 connectors. How can dimension:
Numbers of VA (if adding VA can make some difference)
Number of clusters (perhaps adding more clusters and distributing connectors can help reduce aggregations overlapping)
For each VA: number of processors and memmory
And some final question, is IDN capable of distribute all this heavy work in a fair way?
For 2 VAs with 8GB RAM and 2 core CPU is considered quite a heavy load for 81 sources.
You can initially try sticking to the recommended sizing of VA (16GB RAM and 4 Cores) for both the VAs. This should probably bring down the aggregation time slightly. But adding a new VA(s) with recommended sizing should further help you.
One way to optimize further would be to try and categorize the sources based on regions/zone, identify the category with the highest load and have a cluster with a VAs deployed in that specific region/zone
Thanks @gvscdeep1 VAs sizing increase is on the way now, also we asked additionl VA. My concern is about Tenant, how it is measured? Because after connectors brings all accounts and entitlements data into Tenant, then proccess is doing there, not in VAs.
Tenant sizing is a more complex topic. Remember there’s quite a bit of shared resources and microservices that also have some ability to scale with demand. But SailPoint establishes some of the initial resources based off of your licensing and they have operational teams to monitor and manage from there.
Hi Alex and thanks for responding. This is just my question, documentation does not say anything about the multiple variables to take into account, in order to size the Tenant.
I know that Sailpoint sells solution with an estimative size according to pre-selling assesment. But after the onboarding, how can I measure if a source integration will fit with actual size, or I have to ask client to buy more resources?
This just happened with a source integration, on which the actual tenant size is not enough, and all tenant is very slow (several actions were stuck). So we have to disable new source and now are workinf with Sailpoint on incrrasing tenant size.
Bug I would like to have som parameters, in order to ensure client that the tenant and VAs hace sufficients resourcew to work fine, not to verify after roll out.