During refresh. Not authorized to perform ... Storage account: SASKEY

[dbeavon3]

A number of datasets are failing with a strange error - not authorized to perform this operation using this permission. Storage account: SASKEY

Here is the whole error .

This request is
not authorized to perform this operation using this permission. Storage account: SASKEY

Storage account: SASKEY https://pbipprodeastusdata7.blob.core.windows.net', RequestID: '0358ef29-f01e-003f-3b9a-d892bc000xxx', HttpStatusCode: '403''
Cluster URI: WABI-WEST-US-C-PRIMARY-redirect.analysis.windows.net

Can someone tell me why the PBI service has been failing so bigly these past two weeks?  This sort of crap just keeps flying in every direction and I have opened a number of support cases.  I don't know how Microsoft engineers can be so disinterested in the quality of their platform. 

Please let me know if I can expect to keep encountering this one.

[Anonymous]

Hi @dbeavon3 

You can refer to the link inside the fourth reply. Microsoft Fabric Support and Status | Microsoft Fabric

Usually when a large-scale or generalized problem occurs, our engineers will announce the problem on the Support page and give the appropriate progress or alternative solutions. When the issue is fully fixed, there will be a corresponding alert. So if you have a generalized problem, you can refer to this Support page. 

You can also refer to this link below, which documents some of the known issues, hopefully it will help you.

Microsoft Fabric Known Issues

Best Regards,
Community Support Team _ Ailsa Tao
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[dbeavon3]

Hi @Anonymous 
Are you a Microsoft FTE?  I'm working with support people on my outage and it was widespread across our datasets, and I'm assuming for other customers as well.  There was nothing about this ("SASKEY bug") that would be specific to my environment.

I do not  have any confidence that engineers at Microsoft ever announce "large-scale or generalized problems".  I have opened LOTS of cases over the past two weeks, and in not a single instance did they refer me to a prior announcement.  (.. despite the fact that the mindtree-CSS engineer was already fully aware of the outage and told me details about it within less than an hour of opening my case).

IMHO This is not an efficient way to handle outages.  Mindtree engineers shouldn't need to take dozens of cases about well-known PBI outages, any more than I want to create dozens of cases about well-known outages.  Microsoft is not motivated to be transparent or proactive in this area.  The only way this is ever going to be fixed is if customers take responsibility for demanding an announcement (ie. the first customers impaced would demand an announcement, and other customers get to benefit from it, by not repeating their own investigations or opening their own support tickets).

Do you know if there is a mechanism for customers to demand an announcement in the public about our outage?  This is the most BASIC level of customer service I can think of from a support organization, and I can use the public announcement to communicate to the co-workers at my organization who are also impacted.

[Anonymous]

Hi @dbeavon3 

This issue has been fixed . Please verify this on your Service. Thank you for your cooperation !

Best Regards,
Community Support Team _ Ailsa Tao
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[dbeavon3]

Thanks for the update @Anonymous .

I'm looking for better communication from this PBI SaaS product team.  When customers have massive outages, WITHOUT also getting communication from Microsoft, then it is we who will inevitably take the blame .

You say it has been fixed.  What is your source?  Can you share links to information from the PG?

It is unacceptable that this community is often the ONLY channel for customers to get informed about these outages.  Whenever we have extensive outages, and prolific failures, we are hanging in the wind.  Microsoft leaves customers to fend for ourselves without any way of knowing what is going wrong.

Please let me know exactly why you believe "this issue has been fixed".  If this information is from the PG, then hopefully you agree that CUSTOMERS should also be getting it first-hand, rather than from users in the community.

[Anonymous]

Hi @dbeavon3 

Fabric Customers with tenants located in Central US region may encounter issues while interacting with the service, currently the service is available in read-only mode. There is an ongoing Azure outage and engineers are investigating the issue, up to date status on the investigation can be found at Azure Service Health

Fabric Customers with tenants located in North Central US J cluster may experience delays while interacting with the service. Engineers are investigating the issue and an update will be provided soon.

Microsoft Fabric Support and Status | Microsoft Fabric

Best Regards,
Community Support Team _ Ailsa Tao
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[dbeavon3 · Accepted solution]

Here is a communication I finally received about the SASKEY outage in our region.  I received this today (7/29).

SUMMARY OF IMPACT:
Between 21:00 UTC - and 23:20 UTC on 17 July 2024, you have been identified as a customer using Microsoft Power BI who may experience issue rendering reports, refreshing & publishing datasets. 

Root cause will be provided generally in 14 days.

... I just love how these communications are always point-to-point emails ("you have been identified"). 
😅

It is so wasteful for every impacted customer of Power BI to investigate these outages independently.  It requires that we contact Mindtree, and wait two weeks for a notification about the outage.  The eventual notification from Microsoft is a reactive deliverable, and it is clearly one that we wouldn't have otherwise received.  I guarrantee that there was telemetry on the day of the outage, and Microsoft could have alerted their customers within just hours of the outage if they were motivated to do that.

I know it is not in the best interest of the PG to be more public and transparent about these outages.  But the alternative is an absolutely terrible customer experience.  (and inevitably impacts our satisfaction with the support team at Mindtree)

[dbeavon3]

The Incident ID is 523673573 for an outage which happened on 17th July.

[Anonymous]

Hi  @dbeavon3 

Even if you go and refresh these semantic models manually, does it still report this error? Is there a current semantic model that can be refreshed properly?

Best Regards,
Community Support Team _ Ailsa Tao
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[dbeavon3]

HI @Anonymous 

Are you a Microsoft FTE?  Can you share about the outage?  Someone at Mindtree tried to call me on the phone, and we got disconnected half-way thru the discussion.  I'm not sure why they always have calls, when it is much better to have written communication.

In any case, before we lost our connection they said there was some sort of an outage.  But I was asking for a reliable reference/link that I can share with my team, since the problem was quite widespread across our environment.  I had already checked the "azure status" page for PBI and of course that never shows any information. Not sure why there is an azure status if it never shows status information or outages.

Any information about outages yesterday would be appreciated.  Maybe it was limited to customers in one region like West.