This issue relates to scaling up the environment that was showcased at the F2F workshop on 19 April. It has a direct dependency on the user access issue:
2
Moreover, it requires the following subtasks:
[] ~a wrapper script that can deploy self-contained resource groups for sets of users in Terraform~
[ ] Instead of a wrapper script, we will create three distinct versions of the deployment with different configurations
[ ] We will start with a single environment, and deploy others to accommodate additional users
[X] investigation of linking a Microsoft ML workspace to the access portal
We investigated the ML workspace linkage determined the following:
We will pre-install OHDSI packages/dependencies in each user-specific compute instance through a startup script (rather than referencing an existing pre-built image)
Much of the ML workspace will need to be manually configured in the Azure portal because the configuration options within Terraform are limited.
This issue relates to scaling up the environment that was showcased at the F2F workshop on 19 April. It has a direct dependency on the user access issue:
2
Moreover, it requires the following subtasks: