-
Notifications
You must be signed in to change notification settings - Fork 556
Open
Labels
priority/backlogIssue is approved and in the backlogIssue is approved and in the backlog
Description
We have a scenario where agents are using azure_imds node attestor but the nodes are AKS k8s nodes. These nodes have ephemeral OS disks that can be wiped/replaced on upgrades.
I have tested different locations on the disks and even in /var/lib the data gets wiped
Proposed solution:
- Add a new key_manager plugin for
azure_key_vault - All agents can share a key vault if needed
- On agent start up kick off a background job that will try to delete any keys that have an expiration older than
max(customTTL, minimumTTLAllowed) - Either move as much of the server azure key vault key_manager plugin code to common or use it to create an equivalent agent package
currently using this implementation to test with
https://github.com/0xRBell/spire/pull/2/files
Metadata
Metadata
Assignees
Labels
priority/backlogIssue is approved and in the backlogIssue is approved and in the backlog