Azure VM Performance Troubleshooting with Performance Diagnostics

Define the symptom

Record when the problem occurs, which applications are affected and whether the issue is CPU, memory, disk or network related. Capture the VM size and operating system before making changes.

Use Azure Monitor signals

Review CPU, disk and network metrics over the incident window. Look for sustained saturation versus short spikes. A single high metric does not establish the root cause.

Run performance diagnostics

Azure performance diagnostics can collect data and provide insights for high CPU, memory and disk usage. Use the resulting report to guide the next test rather than treating it as an automatic fix.

Correlate with Windows counters

For Windows workloads, correlate Azure metrics with Performance Monitor counters and application logs. This can reveal whether high disk latency is caused by a storage workload or an application queue.

Check recent changes

Review VM resize operations, application deployments, Windows updates, scheduled tasks, backup jobs and security scans. A performance incident often starts with a change that happened shortly before the symptom.

Validate the remediation

After a change, monitor the same metrics for a representative period and compare them with the baseline. Document the evidence used to close the incident.

Get-Counter '\Processor(_Total)\% Processor Time'
Get-Counter '\Memory\Available MBytes'
Get-Counter '\LogicalDisk(*)\Avg. Disk sec/Read'