I'm looking to better understand how the turbonomic data model is applied to AWS. So far I have gleaned the following:
1. datacenter entities are effectively AWS regions
2. physical machine entities are AWS availability zones
3. virtual machines are AWS virtual machines ;-)
I'm investigating how to align categories of VMs with a particular category of DC - to simulate data/processing location requirements.
Simplistically, I want to stipulate that CAT A VMs only run in CAT A DCs and that CAT B VMs run in CAT B DCs .
I have a test environment with VMs in two different aws regions (i.e.: two "datacenter"s). I have created custom groups for two types of VMs (CAT A and CAT B) and two types of DCs (CAT A DC and CAT B DC). Using a merge placement policy for the CAT A and CAT B DCs and modifying the action mode policy settings for VM and storage move to manual, I get the expected placement violation reported and the associated move action for both as "pending acceptance". When I accept the action I get an exception status from AWS - I'll need to investigate that, but assume it's an API failure/limitation? - any thoughts from the community are welcome.
A secondary question. What determines which AWS DataCenters are discovered by Turbonomic? My inventory includes some, but not all AWS DCs where I have no VMs deployed. Indeed, if I create VMs in a DC that its not in the inventory, they are not discovered (the DC in question is was-eu-west-2).
I'll keep exploring, but any input would be appreciated.