Azure Batch + Azure Monitor Agent: correlating Computer names with pool names (Python SDK)

Introduction

We're using the Python azure-mgmt-batch SDK to create Azure Batch pools backed by VMSS, with Azure Monitor Agent (AMA) enabled via the extensions field in the ARM pool definition. AMA is shipping Linux performance counters (CPU, memory etc.) to a Log Analytics workspace via a Data Collection Rule.

Here is the relevant pool creation code just for reference:

 pool_params = Pool(
      identity=BatchPoolIdentity(
          type=PoolIdentityType.USER_ASSIGNED,
          user_assigned_identities={
              node_monitoring_identity_resource_id: UserAssignedIdentities()
          },
      ),
      vm_size=self.pool_vm_size,
      deployment_configuration=DeploymentConfiguration(
          virtual_machine_configuration=VirtualMachineConfiguration(
              image_reference=ImageReference(...),
              node_agent_sku_id=sku_to_use,
              extensions=[
                  VMExtension(
                      name="AzureMonitorAgent",
                      publisher="Microsoft.Azure.Monitor",
                      type="AzureMonitorLinuxAgent",
                      type_handler_version="1.0",
                      auto_upgrade_minor_version=True,
                      enable_automatic_upgrade=True,
                      settings={
                          "authentication": {
                              "managedIdentity": {
                                  "identifier-name": "mi_res_id",
                                  "identifier-value": node_monitoring_identity_resource_id,
                              }
                          }
                      },
                  )
              ],
          )
      ),
      ...
  )
  batch_mgmt_client.pool.create(
      resource_group_name=batch_account_resource_group,
      account_name=batch_account_name,
      pool_name=pool_name,
      parameters=pool_params,
  )

  # DCR association created on the pool ARM resource after creation
  monitor_client.data_collection_rule_associations.create(
      resource_uri=created_pool.id,
      association_name="ama-dcr-association",
      body=DataCollectionRuleAssociationProxyOnlyResource(
          data_collection_rule_id=dcr_resource_id,
      ),
  )

The problem:

Querying CPU metrics per node works fine, example KQL query:

Perf
  | where ObjectName == "Processor" and CounterName == "% Processor Time" and InstanceName == "total"
  | summarize avg(CounterValue) by Computer, bin(TimeGenerated, 5m)

Result:

┌─────────────────────────┬──────────────────────┬──────────────────┐
│        Computer         │    TimeGenerated     │ avg_CounterValue │
├─────────────────────────┼──────────────────────┼──────────────────┤
│ <computer_name_a>000000 │ 2026-04-27T23:30:00Z │ 78.4             │ ├─────────────────────────┼──────────────────────┼──────────────────┤
│ <computer_name_a>000001 │ 2026-04-27T23:30:00Z │ 81.2             │
├─────────────────────────┼──────────────────────┼──────────────────┤
│ <computer_name_a>000002 │ 2026-04-27T23:30:00Z │ 75.9             │
├─────────────────────────┼──────────────────────┼──────────────────┤
│ <computer_name_b>000000 │ 2026-04-27T23:30:00Z │ 12.1             │
├─────────────────────────┼──────────────────────┼──────────────────┤
│ <computer_name_b>000001 │ 2026-04-27T23:30:00Z │ 10.8             │
└─────────────────────────┴──────────────────────┴──────────────────┘

I need to be able to correlate these Computer names with the Batch pool name in order to aggregate metrics per pool in Grafana, so that the end user doens't need to double check the computer names. This is especially important for easily querying performance metrics of pools which allocate dozens of nodes that are automatically deleted as soon as the corresponding batch job ends.

There is no field in the Perf table, the _ResourceId, or the Computer name that contains the Batch pool name.
The VMSS UUID and the Computer name prefix are both auto-generated by Azure and have no documented relationship to the pool name.

My initial goal was to inject a custom BatchPoolRegistry_CL table into the same Log Analytics workspace containing the PoolName → Computer mapping, so we can join it with the Perf table.
What I've investigated:

• BatchManagementClient.pool.get() — ARM pool response does not expose the underlying VMSS resource group or UUID
• compute_node.list() (legacy azure-batch SDK) — node IDs follow tvmps_{hex}_p format with no obvious link to the VMSS UUID or Computer name

Questions:

1. Is there any API in azure-mgmt-batch (Python management SDK) that exposes the underlying VMSS resource group or a stable identifier correlating with what AMA reports as Computer?
2. Alternatively, is there an officially supported way to associate a Batch pool name with its AMA-monitored nodes?

Thanks!