| Tags |
Flexible labels, multiple per machine. Use for filtering, commissioning, or deployment selection. Can be manual or automatic (XPath). |
Juju ignores tags directly, but you can target machines indirectly by tag when adding them to models or constraints. |
Same as Juju—tags aren’t first-class in OpenStack, but can help when mapping hardware to specific use cases before handoff. |
No direct tie-in to RBAC; tags are just metadata. |
| Availability zones (AZs) |
One per machine. Conceptually “fault domains.” In plain MAAS, it’s just a label. |
Juju respects AZs as failover zones. Workloads can be scheduled across zones for HA. |
OpenStack also respects AZs as placement/failover domains; ensures workloads don’t all land in the same failure point. |
No RBAC connection—zones don’t control access, just placement. |
| Resource pools |
One per machine. Just a label in plain MAAS, often used for org or project grouping. |
Juju does not use pools. |
OpenStack does not use pools. |
Pools are the main grouping mechanism tied to RBAC. Assigning a pool restricts which users or teams can access those machines. |
