get cpu utilization of virtual machines in azure using python sdk

Question

I am trying to get CPU utilization of virtual machines using azure python APIs. like if a virtual machine has 2 cpus, i need overall utilization (means cpu1+ cpu2).

What are the possible methods to get cpu utilization of virtual machines?

Answer 1

Possibly you could use Azure Monitoring libraries for python to obtain the Percentage CPU metric on Azure VMs. Install azure-mgmt-monitor package and call the list methods in MetricsOperations class

Example - Metrics

import datetime
from azure.mgmt.monitor import MonitorManagementClient

# Get the ARM id of your resource. You might chose to do a "get"
# using the according management or to build the URL directly
# Example for a ARM VM
resource_id = (
    "subscriptions/{}/"
    "resourceGroups/{}/"
    "providers/Microsoft.Compute/virtualMachines/{}"
).format(subscription_id, resource_group_name, vm_name)

# create client
client = MonitorManagementClient(
    credentials,
    subscription_id
)

# You can get the available metrics of this specific resource
for metric in client.metric_definitions.list(resource_id):
    # azure.monitor.models.MetricDefinition
    print("{}: id={}, unit={}".format(
        metric.name.localized_value,
        metric.name.value,
        metric.unit
    ))

# Example of result for a VM:
# Percentage CPU: id=Percentage CPU, unit=Unit.percent
# Network In: id=Network In, unit=Unit.bytes
# Network Out: id=Network Out, unit=Unit.bytes
# Disk Read Bytes: id=Disk Read Bytes, unit=Unit.bytes
# Disk Write Bytes: id=Disk Write Bytes, unit=Unit.bytes
# Disk Read Operations/Sec: id=Disk Read Operations/Sec, unit=Unit.count_per_second
# Disk Write Operations/Sec: id=Disk Write Operations/Sec, unit=Unit.count_per_second

# Get CPU total of yesterday for this VM, by hour

today = datetime.datetime.now().date()
yesterday = today - datetime.timedelta(days=1)

metrics_data = client.metrics.list(
    resource_id,
    timespan="{}/{}".format(yesterday, today),
    interval='PT1H',
    metric='Percentage CPU',
    aggregation='Total'
)

for item in metrics_data.value:
    # azure.mgmt.monitor.models.Metric
    print("{} ({})".format(item.name.localized_value, item.unit.name))
    for timeserie in item.timeseries:
        for data in timeserie.data:
            # azure.mgmt.monitor.models.MetricData
            print("{}: {}".format(data.time_stamp, data.total))

# Example of result:
# Percentage CPU (percent)
# 2016-11-16 00:00:00+00:00: 72.0
# 2016-11-16 01:00:00+00:00: 90.59
# 2016-11-16 02:00:00+00:00: 60.58
# 2016-11-16 03:00:00+00:00: 65.78
# 2016-11-16 04:00:00+00:00: 43.96
# 2016-11-16 05:00:00+00:00: 43.96
# 2016-11-16 06:00:00+00:00: 114.9
# 2016-11-16 07:00:00+00:00: 45.4