Unplanned host degradation caused disk corruption and data loss - how to get RCA and SLA credits?

Question

Unplanned host degradation caused disk corruption and data loss - how to get RCA and SLA credits?

Al Mokhlif Oud 0

Hello,

A critical incident occurred on Wednesday, February 18, 2026, affecting my virtual machine running Ubuntu in the West/Central Europe region (I'll share subscription details privately if needed).

Timeline of events (all times GMT+3):

4:21 PM: "Degraded performance detected (Unplanned)" — Azure detected that the physical host running my VM was potentially degraded and advised migration before 3/20/2026.
7:05 PM: "Unavailable (Unplanned)" — VM went down for the first time. The OS began experiencing widespread Input/Output errors, unable to spawn any processes (supervisord, sysstat, php services all failed with I/O errors).
10:20 PM: "Unavailable (Unplanned)" — VM went down a second time.
10:52 PM: I attempted a reboot, which did not resolve the issue due to filesystem corruption caused by the failing host.
11:02 PM: I deallocated the VM.
11:31 PM: I restarted the VM (likely placed on a new host).
11:38 PM: I deallocated again to proceed with a backup restore, as the disk was corrupted beyond recovery.

The host degradation caused disk I/O failures that corrupted the filesystem, making the OS unable to execute any binaries. I was forced to restore from a backup dated 2/15/2026, resulting in approximately 3 days of data loss.

All health events are documented in the VM's Resource Health history.

My questions:

How can I obtain a full Root Cause Analysis (RCA) for the host degradation? Why was the VM not auto-migrated before the disk became corrupted?
How do I request SLA credits for this unplanned platform failure? I am on a Developer support plan and cannot file a direct technical support ticket.
How can I confirm that my VM is now running on healthy hardware?

Any guidance from Microsoft engineers or the community would be appreciated.

Thank you

Jilakara Hemalatha 10,365 Reputation points Microsoft External Staff Moderator

2026-02-19T15:35:30.0933333+00:00

Hello

Thank you for reaching out Q/A. We will look into it.

Could you please check the private message and provide the necessary details
Jilakara Hemalatha 10,365 Reputation points Microsoft External Staff Moderator

2026-02-20T09:22:53.3666667+00:00
Hii Al Mokhlif Oud,

You can request SLA credits for the unplanned outage using the following steps. You do not need any additional support plan—even a developer support plan allows you to create a billing support request.
Steps to Request SLA Credits

Sign in to the Azure Portal using the account affected by downtime.

Navigate to Help + support → Create a support request.

For Issue type, select Billing.

For Problem type, select Refund Request and specify that it is an SLA credit request.

Include the following details in your request:

Subscription & [REDACTED]

Affected service(s) (e.g., Virtual Machines) and region

start/end times (with time zone)

Description of the impact, including any data loss

Review your contact details and click Create to submit the request.

Reference: https://learn.microsoft.com/en-us/partner-center/billing/request-credit
Al Mokhlif Oud 0 Reputation points

2026-02-21T12:34:00.19+00:00
Hello,

Thank you for the guidance on filing the SLA credit request through billing — I will proceed with that.

However, I want to stress the severity of this incident. This was not just downtime — the host degradation caused complete filesystem corruption that destroyed my production environment. If I had not had a backup, my entire server and all its data would have been permanently lost.

I have a few critical concerns that I need addressed:

Root Cause Analysis: What exactly caused the physical host to degrade? The console logs show MANA network adapter failures ("HWC: Request timed out", "Failed to deregister device", worker failures on mana.eth.0 and mana.rdma.1) before the disk I/O errors began. Was the MANA adapter failure a symptom of the host degradation or a separate issue?

Why was the VM not auto-migrated? Azure had 6 hours and multiple warning signs to act:

4:21 PM GMT+3: Host flagged as degraded

7:05 PM GMT+3: VM went unavailable briefly (first crash — a clear sign the host was actively failing)

10:20 PM GMT+3: VM went down permanently, disk corrupted beyond recovery

After the first crash at 7:05 PM, it should have been obvious the host was rapidly deteriorating. Azure had over 3 more hours after that first crash to perform an emergency migration, but nothing was done. Instead, the original notification gave a 30-day deadline (3/20/2026) while the host was actively dying. Why did Azure not perform an emergency migration when the host was clearly failing?

Prevention: What guarantees can Microsoft provide that my VM is now on healthy hardware and that this will not happen again? What proactive monitoring or auto-migration improvements are being made to prevent host failures from causing data corruption?

Escalation: Given the severity of this incident — complete data corruption and permanent data loss caused by Azure's hardware, with 6 hours of ignored warning signs — I would appreciate this case being escalated to a senior engineer or support manager who can provide a thorough investigation.

I rely on Azure for my production workloads and I need confidence that my infrastructure is safe. Please advise.

Thank you

1 answer

Your answer

Jilakara Hemalatha 10,365 Reputation points Microsoft External Staff Moderator

2026-02-19T15:35:30.0933333+00:00

Hello

Thank you for reaching out Q/A. We will look into it.

Could you please check the private message and provide the necessary details
Jilakara Hemalatha 10,365 Reputation points Microsoft External Staff Moderator

2026-02-20T09:22:53.3666667+00:00

Hii Al Mokhlif Oud,

You can request SLA credits for the unplanned outage using the following steps. You do not need any additional support plan—even a developer support plan allows you to create a billing support request.
Steps to Request SLA Credits

Sign in to the Azure Portal using the account affected by downtime.

Navigate to Help + support → Create a support request.

For Issue type, select Billing.

For Problem type, select Refund Request and specify that it is an SLA credit request.

Include the following details in your request:

Subscription & [REDACTED]

Affected service(s) (e.g., Virtual Machines) and region

start/end times (with time zone)

Description of the impact, including any data loss

Review your contact details and click Create to submit the request.

Reference: https://learn.microsoft.com/en-us/partner-center/billing/request-credit

Answer 1

Hello,

Thank you for the guidance on filing the SLA credit request. However, the links you shared describe the process for CSP (Cloud Solution Provider) partners through Partner Center. I am not a CSP partner — I am a direct Azure customer with a Pay-As-You-Go subscription.

I tried to file a billing support request through the Azure portal (Help + Support → Billing → Help with credit request for an outage), but the portal keeps redirecting me to Q&A. It appears a Standard support plan ($100/month) is required to submit a direct support ticket. It does not seem fair to require me to pay $100 just to claim credits for an outage that was caused by Azure's own hardware failure.

Could you please help initiate the SLA credit request from your side, or advise on how I can submit one without upgrading to the Standard plan? I would greatly appreciate any assistance in routing this request to the appropriate billing team internally.

Thank you!

Share via

Unplanned host degradation caused disk corruption and data loss - how to get RCA and SLA credits?

1 answer

Your answer