Azure Virtual Machine Scale Sets

This integration can be enabled via the Connect an Azure Account page.

This integration has 24 metrics.

List of Alerts

List of Dashboards

Azure Virtual Machine Scale Sets

The dashboard provides information on the Azure Virtual Machine Scale Sets integration.

List of Metrics

Monitoring and Troubleshooting Azure Virtual Machine Scale Sets

This document describes important metrics and queries that you can use to monitor and troubleshoot Azure Virtual Machine Scale Sets.

The metrics covered in this document are applicable only to Azure Virtual Machine Scale Sets. Sysdig offers a different integration for Azure Virtual Machines.

Resources

VM Status and Resources

Use the following query to check if the VMs are available. This metric is currently available to only a small set of users.

azure_compute_virtualmachinescalesets_vm_availability_metric_avg != 1

A return value other than 1 indicates a problem.

Percentage CPU Used

Use the following query to get the percentage of CPU used by the VMs:

azure_compute_virtualmachinescalesets_percentage_cpu

Available Memory

Use the following query to get the available memory of the VMs (in bytes):

azure_compute_virtualmachinescalesets_available_memory_bytes

Unfortunately, no metric is available for used memory or maximum memory. Therefore, you cannot calculate the percentage of available memory.

Having 100MB left in a VM with 500MB max and having 100MB left in a VM with 2000MB max are totally different memory calculations.

Network

Network Bytes In

The following query returns the number of bytes received on all the network interfaces by the Virtual Machine (the incoming traffic):

sum_over_time(azure_compute_virtualmachinescalesets_network_in_total[$__interval])

Network Bytes Out

The following query returns the number of bytes out on all the network interfaces by the Virtual Machine (the outgoing traffic):

sum_over_time(azure_compute_virtualmachinescalesets_network_out_total[$__interval])

Disks

The Virtual Machines have an ‘OS’ disk, but they can also have attached Data Disks. The Data Disks have a Logical Unit Number (LUN) which specifies the slot in which the data drive appears when mounted. The valid LUN values are 0 through 15. The ’lun’ label available in the data disk metrics is useful to differentiate which metric comes from which data disk.

OS Disk Byte Rate

Use the following query to get the amount of read and write bytes per second in the OS disk:

azure_compute_virtualmachinescalesets_os_disk_read_bytes_sec

azure_compute_virtualmachinescalesets_os_disk_write_bytes_sec

OS Disk IOPS

Use the following query to get the amount of read and write bytes per second for IOPS:

azure_compute_virtualmachinescalesets_os_disk_read_operations_sec

azure_compute_virtualmachinescalesets_os_disk_write_operations_sec

OS Disk Queue

Use the following query to get the amount of read and write bytes per second for queue depth:

azure_compute_virtualmachinescalesets_os_disk_queue_depth

Data Disk Byte Rate

In a similar fashion, use the following query to get the read and write bytes per second in the data disk:

azure_compute_virtualmachinescalesets_data_disk_read_bytes_sec

azure_compute_virtualmachinescalesets_data_disk_write_bytes_sec

Data Disk IOPS

se the following query to get the read and write bytes per second in the data disk for IOPS:

azure_compute_virtualmachinescalesets_data_disk_read_operations_sec

azure_compute_virtualmachinescalesets_data_disk_write_operations_sec

Data Disk Queue

se the following query to get the read and write bytes per second for queue depth:

azure_compute_virtualmachinescalesets_data_disk_queue_depth

Premium Disks Cache

Premium disks (both OS and data disks) have metrics for their read cache miss and hit, which return a percentage between 0 and 100. Ideally the hit percentage should be 100%, and the miss percentage should be 0%.

Use the following query to get the values in OS disk:

azure_compute_virtualmachinescalesets_premium_os_disk_cache_read_miss

azure_compute_virtualmachinescalesets_premium_os_disk_cache_read_hit

Use the following to get the values in the data disks:

azure_compute_virtualmachinescalesets_premium_data_disk_cache_read_miss

azure_compute_virtualmachinescalesets_premium_data_disk_cache_read_hit

VM Cached and Uncached IOPS Consumed

Use the following query to get the cached and uncached VM IOPS consumed:

azure_compute_virtualmachinescalesets_vm_cached_iops_consumed_percentage

azure_compute_virtualmachinescalesets_vm_uncached_iops_consumed_percentage

A value near 100% indicates a bottleneck. This occurs when the VM machine IOPS reaches its limit. In that case, the bottleneck issue can be fixed by increasing the size of the VM. Before fixing it you should check the IOPS of the attached disks. If their IOPS are also at 100%, the bottleneck is in the disks.

Credits

CPU Credits

The following metric, only available on B-series burstable VMs, returns the remaining CPU credits available to burst:

azure_compute_virtualmachinescalesets_cpu_credits_remaining

Disk Credits

OS Disk Used Burst Bandwidth and IO Credits

Use the following query to get the percentage of used burst bps (bytes per second) credits for the OS disk:

azure_compute_virtualmachinescalesets_os_disk_used_burst_bps_credits_percentage

Use the following query to get the percentage of used burst IO credits for the OS disk:

azure_compute_virtualmachinescalesets_os_disk_used_burst_io_credits_percentage

Data Disk Used Burst Bandwidth and IO Credits

Use the following query to get the percentage of used burst bps (bytes per second) credits for the data disks:

azure_compute_virtualmachinescalesets_data_disk_used_burst_bps_credits_percentage

Use the following query to get the percentage of used burst IO credits in the data disks:

azure_compute_virtualmachinescalesets_data_disk_used_burst_io_credits_percentage