Open darkn3rd opened 2 years ago
Hi @darkn3rd, thank you for raising this issue, your issue description is stellar!
Would you like to open a PR with the fix?
Thanks!
Thanks for the feedback. I used to work in QA back in a time when customers were not the testers as per industry norms these days. In those days, we'd get lectured for bad issue writeups. Is anyone able to update and make changes, or do you need help with this?
Segue, is there a way to run CI over example snippets to more rapidly test, as such examples require some effort to test on the perspective platforms? If not, I could make another enhancement request for doing an examples directory, we can dump reusable snippets in there, that could later be forklifted back into the docs. This way they could be tested downstream. Currently, I do ad-hoc tests using the free tier/evals on perspective cloud (Azure, GCP, AWS).
You can assign this to me, I will work on this when I do Azure experiments again.
When documenting steps to setup a managed identity, the documentation incorrectly instructs the reader to essentially the use a
principalId
, which is different thanprincipal_id
. The principal's object id akaclientId
should be used instead.The current incorrect documentation is:
The correct documentation should be:
Similarly the Terraform code looks incorrect, and it should be:
The azure_rm provider is quite confusing because the return value (or rather attribute) from
azurerm_user_assigned_identity
shares the same name as argument forazurerm_role_assignment
, but they in fact represent different things.References:
For the differences between these two in Azure:
The Terraform documentation is confusing given the name:
Given the description, the principal is the managed identity that was created, and the ID of that principal is the
client_id
, so in some way,principal_id
could be a shortened version ofprincipal_client_id
.