Unanswered

When an event is added, updated or deleted (V3) "Bad Gatway" Error "Request failed. Try again later."

(0) Share

Report

Posted on by ericwhitley

137

We have a flow that uses the When an event is added, updated or deleted (V3) connector that points to a shared calendar.

The shared calendar receives thousands of invitations per day, and approximately every 3 to 4 weeks it generates a "Bad Gateway" error, with message "Request failed. Try again later.":

"body": {

"error": {

"code": 502,

"source": "flow-apim-unitedstates-002-westus-01.azure-apim.net",

"clientRequestId": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",

"message": "BadGateway",

"innerError": {

"status": 502,

"message": "Request failed. Try again later.\r\nclientRequestId: xxxxxxxxxxxxxxxxxxxxxxxxxxxx\r\nserviceRequestId: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",

"error": {

"message": "Request failed. Try again later."

"source": "office365-scus.azconn-scus-001.p.azurewebsites.net"

}

Once this error occurs, all subsequent triggers fail until the connector is removed & recreated. The errors usually happen overnight and remain until somebody recreates the connector the next morning. This causes us to miss out on thousands of updates from the calendar. (The Analytics from our latest error show over 1000 errored triggers.)

We have noticed that prior to the first Bad Gateway error, the flow receives thousands of triggers within a very short timeframe (1 to 2 minutes.) Usually these are ActionType = 'deleted'. My assumption is that Outlook/Exchange is doing a purge of old data.

I believe the real problem is the flow trigger is overloaded and can't handle all the updates.

What can be done to prevent the Bad Gateway errors?

Can they be captured and at least send an email when an error occurs?

I have thought about using a Trigger Condition to exclude ActionType = 'deleted', but in the Settings the Split is Off, which allows multiple updates per trigger. (This is necessary for other parts of the flow to work properly.)

Can someone from Microsoft confirm what's actually happening?

Thank you

Categories:

General topics

I have the same question (0)

All responses (3)

Answers (0)

michalparal222 86 on at

Like (0)

Report
Hi @ericwhitley ,

In case you would not need a immediate trigger when an anything happened to the calendar event, you could try get the logs from Admin center from admin logs, which can be accessed using HTTP.
I use this pretty often, I just have not tried for calendar events yet.

Simple description

Create an enterprise app in Entra
Assign API permission
Grant admin consent
Generate client secret
Using HTTP generate a access token
Register feed/subscription to get the needed logs
In the feed logs are stored in blobs, so need to get the URL of all blobs and then access then one by one.
Then just read the info and use as you need
Also you might need to filter out some items using Filter array

But you need to take into account that the blobs are created with some time delay (few hours).

I can elaborate on this one if it would suits your needs.

Was this reply helpful? Yes No
ericwhitley 137 on at

Like (0)

Report

Michal, thank you for the feedback but we need to get the invitations immediately.

Our current flow works fine most of the time, it just fails every few weeks and I would like to know why and how to prevent that.

Was this reply helpful? Yes No
ericwhitley 137 on at

Like (0)

Report

The flow stopped running again today during work hours with the same "Bad Gateway" error.

Prior to the error we received over 500 triggers to the flow within a few minutes. Almost all of them are ActionType = deleted.

A few hours before the error we received another error, but this didn't cause the flow to stop running: "Too many requests sent to the service. Try again later."

This is new error information:

{
"statusCode": 429,
"headers": {
"Cache-Control": "no-store, no-cache",
"Pragma": "no-cache",
"Retry-After": "32",
"Set-Cookie": "ARRAffinity=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx;Path=/;HttpOnly;Secure;Domain=office365-scus.azconn-scus-001.p.azurewebsites.net,ARRAffinitySameSite=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx;Path=/;HttpOnly;SameSite=None;Secure;Domain=office365-scus.azconn-scus-001.p.azurewebsites.net",
"Strict-Transport-Security": "max-age=31536000; includeSubDomains",
"x-ms-request-id": "5a8a2f13-fd8e-46d7-a4e5-4cde0cd3fb9f",
"X-Content-Type-Options": "nosniff",
"X-Frame-Options": "DENY",
"x-ms-environment-id": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
"Timing-Allow-Origin": "*",
"x-ms-apihub-cached-response": "true",
"x-ms-apihub-obo": "false",
"Date": "Mon, 24 Jun 2024 20:53:02 GMT",
"Content-Length": "370",
"Content-Type": "application/json",
"Expires": "-1"
},
"body": {
"status": 429,
"message": "Too many requests sent to the service. Try again later.\r\nclientRequestId: 941fee4b-3aa6-46d2-a936-12525087aea9\r\nserviceRequestId: 5a8a2f13-fd8e-46d7-a4e5-4cde0cd3fb9f",
"error": {
"message": "Too many requests sent to the service. Try again later."
},
"source": "office365-scus.azconn-scus-001.p.azurewebsites.net"
}
}

I still believe the problem is the flow is overtaxed. How can this be prevented? What are some workarounds?

Thanks

Was this reply helpful? Yes No