I have workerrole on product, but failed to work after a while, it has signalr in it, it returns 503 error. I thought It could be my code, so I published again to the production, still got the same error. Then I published to staging, found it worked. Then I did swap, thought I can swap the working one from staging to production, but still got same error after the swapping. Then I found it is not the code, it is the VIP, the product VIP just returns the 503 error. The VM is still there if someone want to take a look.
So this is a big problem, Azure was not quite stable, I thought we can keep two VM on both production and staging, once the production failed, I can always swap staging over, at least our customer can still access something. But with such VIP problem, I have to update all our DNS to new VIP, and it takes time to update. It is a show stopper.
Please help.