Performance analysis needed on the ANF volume eus-ccdofe-rbmscratchado-prod

Hi Srikanth L,

Thank you for contacting us about the azure ANF volume performance issue. We have begun looking into it and will share our suggestions with you as soon as possible. If you have any additional details or concerns in the meantime, please let us know. We appreciate your patience while we work on this.

Hi Srikanth L,

Thanks for reaching out in Microsoft Q&A forum,

I would like a detailed analysis of the volume’s performance and let us know the root cause of this issue.

The root cause of the slowness on volume and latency spike on April 22–23 for volume eus-ccdofe-rbmscratchado-prod was caused by the workload’s I/O demand exceeding the volume’s provisioned throughput limit. With automatic QoS, throughput is determined by the volume quota and service level (64 MiB/s per TiB on Premium). When demand exceeds this limit, ANF throttles I/O, requests queue up, and latency increases from the normal 1–3 ms to 8–10 ms. The “Throughput limit reached” metric hitting 1 and the QoS latency delta spike (~3.47 sec) confirm this was QoS throttling not a storage hardware, network, or platform issue. The storage system behaved exactly as designed.

To prevent recurrence:

• Increase throughput: For automatic QoS, increase the volume quota (overprovision if needed) or upgrade to Ultra service level. For manual QoS, directly increase the assigned throughput up to 4,500 MiB/s maximum
• Stagger heavy jobs: Avoid running multiple high-IOPS batch jobs simultaneously
• Set early warning alerts: Monitor “Throughput limit reached,” QoS latency delta, and sustained latency >5 ms

Key facts from Microsoft’s official documentation:

• When volume throughput reaches its maximum, response times (latency) increase this is expected QoS behavior, not a performance issue caused by storage
• You can dynamically increase volume quota without interrupting I/O, and the throughput limit changes nearly instantly
• Maximum empirical throughput is 4,500 MiB/s; beyond 70.31 TiB quota on Premium, additional quota doesn’t increase throughput

Reference:

• General performance considerations for Azure NetApp Files | Microsoft Learn
• Metrics for Azure NetApp Files | Microsoft Learn
• Service levels for Azure NetApp Files | Microsoft Learn

Kindly let us know if the above helps or you need further assistance on this issue.

Please “up-vote” wherever the information provided helps you, this can be beneficial to other community members.

I am not convinced with the answer. The analysis based on the fact the volume is in premium tier, where volume is already in ultra tier, I would like though analysis done from the backend.

Hi Srikanth L

Apologises,

Could you please share the requested details via private message?