azure-mgmt-fabric-py – DevOps & Infrastructure skill for AI agents

Azure Fabric Management SDK for Python

Manage Microsoft Fabric capacities and resources programmatically.

Installation

pip install azure-mgmt-fabric
pip install azure-identity

Environment Variables

AZURE_SUBSCRIPTION_ID=<your-subscription-id>
AZURE_RESOURCE_GROUP=<your-resource-group>

Authentication

from azure.identity import DefaultAzureCredential
from azure.mgmt.fabric import FabricMgmtClient
import os

credential = DefaultAzureCredential()
client = FabricMgmtClient(
    credential=credential,
    subscription_id=os.environ["AZURE_SUBSCRIPTION_ID"]
)

Create Fabric Capacity

from azure.mgmt.fabric import FabricMgmtClient
from azure.mgmt.fabric.models import FabricCapacity, FabricCapacityProperties, FabricCapacityAdministration, CapacitySku
from azure.identity import DefaultAzureCredential
import os

credential = DefaultAzureCredential()
client = FabricMgmtClient(
    credential=credential,
    subscription_id=os.environ["AZURE_SUBSCRIPTION_ID"]
)

resource_group = os.environ["AZURE_RESOURCE_GROUP"]
capacity_name = "myfabriccapacity"

capacity = client.fabric_capacities.begin_create_or_update(
    resource_group_name=resource_group,
    capacity_name=capacity_name,
    resource=FabricCapacity(
        location="eastus",
        sku=CapacitySku(
            name="F2",  # Fabric SKU
            tier="Fabric"
        ),
        properties=FabricCapacityProperties(
            administration=FabricCapacityAdministration(
                members=["[email protected]"]
            )
        )
    )
).result()

print(f"Capacity created: {capacity.name}")

Get Capacity Details

capacity = client.fabric_capacities.get(
    resource_group_name=resource_group,
    capacity_name=capacity_name
)

print(f"Capacity: {capacity.name}")
print(f"SKU: {capacity.sku.name}")
print(f"State: {capacity.properties.state}")
print(f"Location: {capacity.location}")

List Capacities in Resource Group

capacities = client.fabric_capacities.list_by_resource_group(
    resource_group_name=resource_group
)

for capacity in capacities:
    print(f"Capacity: {capacity.name} - SKU: {capacity.sku.name}")

List All Capacities in Subscription

all_capacities = client.fabric_capacities.list_by_subscription()

for capacity in all_capacities:
    print(f"Capacity: {capacity.name} in {capacity.location}")

Update Capacity

from azure.mgmt.fabric.models import FabricCapacityUpdate, CapacitySku

updated = client.fabric_capacities.begin_update(
    resource_group_name=resource_group,
    capacity_name=capacity_name,
    properties=FabricCapacityUpdate(
        sku=CapacitySku(
            name="F4",  # Scale up
            tier="Fabric"
        ),
        tags={"environment": "production"}
    )
).result()

print(f"Updated SKU: {updated.sku.name}")

Suspend Capacity

Suspend only an approved capacity after assessing active workloads; verify the actual billing and commitment terms rather than promising all charges stop:

client.fabric_capacities.begin_suspend(
    resource_group_name=resource_group,
    capacity_name=capacity_name
).result()

print("Capacity suspended")

Resume Capacity

Resume a paused capacity:

client.fabric_capacities.begin_resume(
    resource_group_name=resource_group,
    capacity_name=capacity_name
).result()

print("Capacity resumed")

Delete Capacity

client.fabric_capacities.begin_delete(
    resource_group_name=resource_group,
    capacity_name=capacity_name
).result()

print("Capacity deleted")

Check Name Availability

from azure.mgmt.fabric.models import CheckNameAvailabilityRequest

result = client.fabric_capacities.check_name_availability(
    location="eastus",
    body=CheckNameAvailabilityRequest(
        name="my-new-capacity",
        type="Microsoft.Fabric/capacities"
    )
)

if result.name_available:
    print("Name is available")
else:
    print(f"Name not available: {result.reason}")

List Available SKUs

skus = client.fabric_capacities.list_skus(
    resource_group_name=resource_group,
    capacity_name=capacity_name
)

for sku in skus:
    print(f"SKU: {sku.name} - Tier: {sku.tier}")

Client Operations

| Operation | Method |

|-----------|--------|

| client.fabric_capacities | Capacity CRUD operations |

| client.operations | List available operations |

Fabric SKUs

| SKU | Description | CUs |

|-----|-------------|-----|

| F2 | Entry level | 2 Capacity Units |

| F4 | Small | 4 Capacity Units |

| F8 | Medium | 8 Capacity Units |

| F16 | Large | 16 Capacity Units |

| F32 | X-Large | 32 Capacity Units |

| F64 | 2X-Large | 64 Capacity Units |

| F128 | 4X-Large | 128 Capacity Units |

| F256 | 8X-Large | 256 Capacity Units |

| F512 | 16X-Large | 512 Capacity Units |

| F1024 | 32X-Large | 1024 Capacity Units |

| F2048 | 64X-Large | 2048 Capacity Units |

Capacity States

| State | Description |

|-------|-------------|

| Active | Capacity is running |

| Paused | Capacity is suspended (no billing) |

| Provisioning | Being created |

| Updating | Being modified |

| Deleting | Being removed |

| Failed | Operation failed |

Long-Running Operations

All mutating operations are long-running (LRO). Use .result() to wait:

# Synchronous wait
capacity = client.fabric_capacities.begin_create_or_update(...).result()

# Or poll manually
poller = client.fabric_capacities.begin_create_or_update(...)
while not poller.done():
    print(f"Status: {poller.status()}")
    time.sleep(5)  # import time before this optional manual polling fragment
capacity = poller.result()

Best Practices

1. Use DefaultAzureCredential for authentication

2. Suspend unused capacities to reduce costs

3. Start with smaller SKUs and scale up as needed

4. Use tags for cost tracking and organization

5. Check name availability before creating capacities

6. Handle LRO properly — don't assume immediate completion

7. Set up capacity admins — specify users who can manage workspaces

8. Monitor capacity usage via Azure Monitor metrics

When to Use

This skill is applicable to execute the workflow or actions described in the overview.

Limitations

• Use this skill only when the task clearly matches the scope described above.
• Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
• Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.