how to fix the VM uncached bandwith issue that is getting triggered again and again

Question

how to fix the VM uncached bandwith issue that is getting triggered again and again

MEHRA, ANCHAL 0

how to fix the VM uncached bandwith issue that is getting triggered again and again due to which our frontend application is not getting loaded, our VM size is Standard E8as v4 (8 vcpus, 64 GiB memory) and our /u02 disk size is Premium SSD LRS 10240 18000

Ankit Yadav 6,345 Reputation points Microsoft External Staff Moderator

2025-12-05T11:36:00.02+00:00
Hello @MEHRA, ANCHAL ,

Based on your description this seems more like your VM is not able to meet the performance which could be due to your application requirement for more resources with the VM but possible due to VM/disk usage you are not getting optimal performance.

Potential short-term fixes that you can try :

Change disk caching to "ReadOnly" if mostly read operations

Temporarily scale up to E16s v4 (384 MB/s uncached) to confirm if it improves the performance.

Long-term solutions for your VM:

Upgrade to E16s v4 for higher uncached limits

Consider Ultra Disks for guaranteed high throughput with low latency

Implement application-level caching (Redis, Memcached)

Configuration optimizations: -> Implement filesystem tuning appropriate for your workload -> Move logs/temp files to separate disk -> Adjust application logging levels to reduce disk writes

Do let me know if these help out to fix the issue for you.
Ankit Yadav 6,345 Reputation points Microsoft External Staff Moderator

2025-12-05T11:36:48.1733333+00:00
If these solutions don't help to fix the issue for you, can you please below queries to help us understand more about your issue.

When did this issue start occurring? Is it happening at specific times (peak hours, scheduled jobs, backups) or all the time ?

Did you recently deploy any application changes or observed any increase user traffic?

What are your disk read/write metrics showing during the incidents? (Check Azure Metrics for your VM and look for "Disk Read/Written Bytes/sec" and "Disk Read/Write Operations/sec)

What percentage of your disk IOPS and throughput are you utilizing when the issue occurs?

How is your /u02 partition formatted and mounted (ext4/xfs, mount options)?

Are you using swap space that might be causing additional disk I/O?

Is your application writing excessive logs or temporary files?

Are other applications/services on this VM competing for disk I/O?

1 answer

Your answer

Ankit Yadav 6,345 Reputation points Microsoft External Staff Moderator

2025-12-05T11:36:00.02+00:00

Hello @MEHRA, ANCHAL ,

Based on your description this seems more like your VM is not able to meet the performance which could be due to your application requirement for more resources with the VM but possible due to VM/disk usage you are not getting optimal performance.

Potential short-term fixes that you can try :

Change disk caching to "ReadOnly" if mostly read operations

Temporarily scale up to E16s v4 (384 MB/s uncached) to confirm if it improves the performance.

Long-term solutions for your VM:

Upgrade to E16s v4 for higher uncached limits

Consider Ultra Disks for guaranteed high throughput with low latency

Implement application-level caching (Redis, Memcached)

Configuration optimizations: -> Implement filesystem tuning appropriate for your workload -> Move logs/temp files to separate disk -> Adjust application logging levels to reduce disk writes

Do let me know if these help out to fix the issue for you.
Ankit Yadav 6,345 Reputation points Microsoft External Staff Moderator

2025-12-05T11:36:48.1733333+00:00

If these solutions don't help to fix the issue for you, can you please below queries to help us understand more about your issue.

When did this issue start occurring? Is it happening at specific times (peak hours, scheduled jobs, backups) or all the time ?

Did you recently deploy any application changes or observed any increase user traffic?

What are your disk read/write metrics showing during the incidents? (Check Azure Metrics for your VM and look for "Disk Read/Written Bytes/sec" and "Disk Read/Write Operations/sec)

What percentage of your disk IOPS and throughput are you utilizing when the issue occurs?

How is your /u02 partition formatted and mounted (ext4/xfs, mount options)?

Are you using swap space that might be causing additional disk I/O?

Is your application writing excessive logs or temporary files?

Are other applications/services on this VM competing for disk I/O?

Answer 1

To address the VM uncached bandwidth issue that is affecting your frontend application, consider the following steps:

Monitor Performance Metrics: Use Azure Monitor to check the performance metrics of your VM, especially focusing on the VM Cached IOPS, Uncached IOPS, and bandwidth consumed. This will help you identify if you're hitting the limits of your VM's capabilities.
Optimize Disk Configuration: Ensure that your disk configuration aligns with your workload requirements. Since you are using a Premium SSD, verify that the disk's performance capabilities are not being exceeded. You may need to consider resizing your disk or using multiple disks if necessary.
Scale Up the VM: If your application requires more resources than your current VM size can provide, consider scaling up to a larger VM size that offers higher uncached bandwidth and IOPS. The Standard E8as v4 may not be sufficient for your workload, especially if the application is latency-sensitive.
Check for Throttling: Investigate if there is any throttling occurring at the VM or disk level. You can refer to the Azure documentation for metrics that indicate whether your VM is being throttled due to exceeding IOPS or bandwidth limits.
Load Balancing: If your application can be distributed across multiple VMs, consider implementing load balancing to distribute the workload evenly, which can help mitigate bandwidth issues.
Review Application Performance: Sometimes, the issue may not solely be with the VM or disk but could also stem from how the application is designed. Review the application code and database queries to ensure they are optimized for performance.
Use Azure Premium Storage: If you are not already using Azure Premium Storage, migrating to it can provide better performance for your VM workloads, especially for I/O-intensive applications.

By following these steps, you should be able to address the uncached bandwidth issues and improve the performance of your frontend application.

References:

Share via

how to fix the VM uncached bandwith issue that is getting triggered again and again

1 answer

Your answer