Hi,
I know this is a rather complex setup and there are ways to make it easer, but I’m rather curios if this could work at all.
Let’s assume I have to following structure:
/ns1
|- approles/automation_role
| - token/automation_role(?)
| - ns2
| | - secrets/
| | | - A
| | | - B
| | - policies/
| | - policy4A
| | - policy4B
| - ns3
| - secrets/
| | - X
| | - Y
| - policies/
| | - policy4X
| | - policy4Y
The basic idea is: Dedicated admins of ns2 + ns3 manage the policies for their secrets independent. Even from the parent namespace. Now, an automation tool needs to read a secret from one of the namespaces and to prevent accidential leakage, it should only be able to perform the operations as specified in the corresponding policy. E.g.: It needs to work with secret “A”, only operations specified in policy “policy4A” must be permitted.
I think this should be possible by creating a token role which has all the policies associated and then creating an orphan token with just the necessary policy.
But where to put the token role and identities? Can this be implemented in a way that the parent namespace doesn’t need to know about all the policies?
Thanks in advance for your help,
Andreas