2
votes

I have an Azure App Service that has a stage slot just for deployment. I got setted 100% traffic to the production slot. But sometimes randomly my production traffic got redirected to the stage slot and all my users went down. After a few minutes the traffic goes back to production slot and everything goes back to normal.

IMG: Requests at stage slot. My last deploy (and stage use) was 1 day before that .

Has anyone been trough something like this?

1
are you sure they are real users and it's only a "ping" request? - Thiago Custodio
@ThiagoCustodio they are, cause we got massive reports at those exact time. And no errors on prod API. - Matheus Oliveira
The only possible explanation to this is a CI/CD that deploys to your deployment slot and after it gets "warmed", it swap the slots. Check if you had deployments near the time you received those requests. - Thiago Custodio
I would raise ticket with Microsoft. - Sadiq Khoja
@ThiagoCustodio, our last deploy to that stage slot was yesterday. And a few hours after that I swapped envs and that's it. Nothing was done today, when that happened. :// - Matheus Oliveira

1 Answers

0
votes

You can try a couple of things to isolate the issue and re-verify the following configuration:

  1. Has routing Production traffic to the Staging slot been configured either automatically from the Deployment slot blade or manually by adding the ‘x-ms-routing-name’ query parameter in the code?

  2. To confirm that the requests are truly not being routed to the Web App in the Staging slot, kindly configure ‘Failed request tracing’ to validate from the application side that the requests are truly not being routed to the Web App in the Staging slot: https://docs.microsoft.com/azure/app-service/troubleshoot-diagnostic-logs#log-detailed-errors

  3. Just to narrow-down the issue, If feasible, restart the WebApp and analyze to see how long the slot does not receive any traffic? When does the WebApp in the slot starts receiving the requests again? Review and capture the timeframe.