Export email step suddenly failing with error - bad gateway, mailbox move in progress

[kgregg]

I have several flows built that look for new emails coming into an O365 shared inbox.  The first step is to export the email which had been running successfully for some time but starting today, all of the flows are erroring out with the following error.  Any ideas?

{

    "statusCode": 502,

    "headers": {

        "Transfer-Encoding": "chunked",

        "Retry-After": "15",

        "Vary": "Accept-Encoding",

        "Strict-Transport-Security": "max-age=31536000",

        "request-id": "de015b54-08b3-4ff7-9c41-3d29f9f52536",

        "client-request-id": "de015b54-08b3-4ff7-9c41-3d29f9f52536",

        "x-ms-ags-diagnostic": "{\"ServerInfo\":{\"DataCenter\":\"West US\",\"Slice\":\"E\",\"Ring\":\"5\",\"ScaleUnit\":\"002\",\"RoleInstance\":\"BY1PEPF00002EF9\"}}",

        "Timing-Allow-Origin": "*",

        "x-ms-apihub-cached-response": "true",

        "Cache-Control": "private",

        "Date": "Wed, 21 Apr 2021 13:27:10 GMT",

        "Content-Type": "application/json",

        "Content-Length": "436"

    },

    "body": {

        "error": {

            "code": 503,

            "source": "flow-apim-msmanaged-na-westus2-01.azure-apim.net",

            "clientRequestId": "4858d938-011e-4b97-a025-d7fb94d1be35",

            "message": "BadGateway",

            "innerError": {

                "error": {

                    "code": "ErrorMailboxMoveInProgress",

                    "message": "Mailbox move in progress. Try again later., The user and the mailbox are in different Active Directory sites."

                }

            }

        }

    }

}

[kierian]

are these mailbox getting migrated 

essage": "Mailbox move in progress. Try again later., The user and the mailbox are in different Active Directory sites."

[kgregg]

Not that I am aware of.  Would we be notified if they were?

[kierian]

Hello 

You will need to check with your internal IT department for this if they doing any mailbox migration 

Have a good day.

[kgregg]

I will check.  These are all cloud (exchange online) mailboxes so there is no migration that we would be doing.  The connection associated with these particular steps is a different user than those mailboxes.  Could anything be occurring on the azure side that would be affected by that?

[kgregg]

Would Microsoft ever have the need to migrate these around in the cloud (e.g. when there are service issues, etc)

[kierian]

possible that something going on with this account on azure side 

if you are global admin and have access to these mailboxes try to create connector under your account and test

[kgregg]

The all started failing right around 8:40 AM EST today

[kierian]

please check with office 365 admin and if need open case with 365 support to confirm nothing been done on this affected user 

[TomasZubov]

@kierian we experince same behaviour on our tenant and there is no migration or nothing in place on Exchange

[Peter1234]

Having same problem but in a logic app.

Export Email from Office 365 Outlook connector is failing with the exact same message.

No migration happening.

This needs to be fixed urgently

[bermo]

Same problem with Peter1234

[kgregg]

This is now going on day 3 for us.  The first day, we figured out that creating a new user and connection for those steps got all of the flows working again but the following day at almost the exact same time, they all started failing again with the same badgateway error.  Tried to create a new user/connection for every flow as a temporary band-aid but only some worked.  The ones that did work have now started failing again this morning.  I put in an incident with Microsoft and worked with them over the past 2 days but they have not come up with any explanation, fix, or workaround as yet.  This is very frustrating.  This was really our first serious buildout of flows for our organization.  We had just gotten these rolled out to our live production processes a couple months ago and now have to tell the business we can't fix them.  How can we continue to build out critical functionality for the business on this platform when issues like this repeatedly happen?

[Peter1234]

Fully agree!

For months this worked perfectly fine and on the 20 April between 09h39 AM and 9h59 AM, something must have changed, because up until 9h39AM all was working well and after 9h59 it was no more possible to export any emails.

[Peter1234]

@kgregg : my workflow is using a trigger "when a mail arrives on a shared mailbox (V2)", yours?

[kgregg]

@Peter1234 yes, that is the connector that is failing out for me.  A couple things to note this morning - there seemed to be a blip in the ones I created yesterday and I got the badgateway error and then at least one of the flows resumed working again.  These are all fairly low volume inboxes with the exception of one that gets probably one email every couple seconds.  That one for a very brief moment this morning threw a throttling error '

Application is over its MailboxConcurrency limit.' but then quickly went back to the badgateway error.

[kgregg]

@They did send out a couple advisories last night - MX252190 and MM252194

Preliminary Root Cause: A portion of the infrastructure responsible for data calls to SharePoint and Microsoft 365 had become unhealthy. Next Steps: We're analyzing performance data and trends on the affected systems to prevent this problem from happening again.

[Peter1234]

@kgregg for me the mailbox is not that heavily used: 10 to 100 a day max. Not a lot of mails, but the functionality is still very important for us.

[kierian · Accepted solution]

In order to resolve the issue please follow the following steps:

1. Confirm the error message and that user is using the “Export  Email (V2)” action
2. Confirm with the user that they are using a Shared Mailbox, if yes then proceed with the next step
3. Have the user populate the Original Email Address field with the shared mailbox they are exporting the emails from

[Peter1234]

@kierian I confirm this resolves the problem for me. Why was the property added without any backwards compatibility or warning to the users of the connector. This should be possible.