Sudden increase in latency when connecting from GCP to Azure

[flagbug]

Hey folks, about 36 hours ago we suddenly saw a 10x increase in latency (from about ~4ms to ~40ms) when connecting from our services running on Google Compute Engine (us-east4) to Azure SQL database and Azure Storage (East US). Is someone else experiencing a similar issue? Services running inside of Azure looks fine, as well as latency from Google Cloud to other services that are not in Azure, so I'm assuming it's something in the networking between the both, but not sure on which side the problem lies.

[Marcel2]

Hello, no known issues reported in GCP side by that date that could have triggered latency or connectivity low performance between GCP us-east4 and Azure SQL/Azure Storage. Additional information would help to narrow down the issue:

- Non us-east4 Compute Engine resources were experiencing similar behavior to same destination?

- Is it resolve now? How much time did it last?

- Have you checked with Azure to double check if they noticed any related behavior?

- As the issue is/was noticed from specific VMs (as per my understanding), do you see any high operation CPU/Memory/Disc by that time? Any error also in logging?

This information may help to pinpoint a bit more this noticed behavior.

[rajdeep-th]

Hi Flagbug,

Can you please help me understand how this issue was solved for you? 

As of now, I am also experiencing a similar issue where I have a database on GCP and my server is on Microsoft Azure and the latency has increased by 10x for the connection between the two. 

This is currently hampering my business and this is the only issue I could come across which was similar to the one I am facing. Your reply on this will be of a great help.

Looking forward for a quick response.

[flagbug]

After dealing with this for a few days, Azure seems to have fixed some faulty configuration or hardware in their datacenter and the issue suddenly disappeared. Unfortunately I don't have a better answer.

[rajdeep-th]

Thanks for the quick response. 

It seems to be the same in my case as well. Today the issue seems to be auto-fixed. It might be due to some fixes done by the Microsoft team from their side.