Azure "successfully starts" deployment, but nothing deployed - azure

Question

After performing numerous deployments of some software to Azure, I'm hitting a strange problem which is stopping the deployment from working. (I am doing all these deployments in this case by using Visual Studio's Package command, and then using the Azure portal's Upload button.)
The portal initially says it has successfully started the deployment:
It also says it is creating the staging deployment:
(ignore timestamp, screenshot is from another attempt):
But that's all. It never proceeds to show the instance, going through various states and finally running, as has always happened before. There are no further notifications and no error messages. (Even after 24hrs, for the avoidance of doubt.)
[2

Answer 1

You can also look at the Deployments log for a Resource Group in the Azure Portal by going to that Resource Group > Deployments. All deployments, successful and failed, should be listed there. If you click a failed deployment there should be a note at the top that says Failed. Click here for details. This often gives some more details to why the deployment failed.
Microsofts documentation for this: Troubleshoot common Azure deployment errors with Azure Resource Manager

Answer 2

This held us up for several days. Incredible as it sounds, there's apparently a major bug in the "new portal" that stops the error reporting from working, resulting in a silent failure instead of an explanatory message or logging. In our case we had simply hit our limit of 20 cores in the Azure subscription, but the portal wasn't letting on.
As soon as we deleted a service we didn't need any more, the deployment worked as normal.
The cause was discovered by pure chance, when someone else tried to create a new cloud service just at the time this problem was occurring. A UI message informed them the new service could not be created due to hitting the limit.
The lack of an equivalent message when updating an existing service is is a jaw-dropping defect in the "new portal". Very sadly we are well used to MS error messages being unhelpful, and often even very misleading, but with this silent failure MS seem to have excelled even themselves as far as standards of error reporting go.
EDIT: the old portal helpfully reports core utilization on the dashboard:
Unfortunately that doesn't seem to work on the new portal.
Of far greater significance, though, is that the OLD portal DOES report the deployment failure, rather than failing silently as the "new portal" does:
...which leads to:
So the moral of the tale seems to be: if you have an unexplained deployment problem, use the old portal (https://manage.windowsazure.com). You will probably then find the cause straight away, as the old portal actually reports the failure reason instead of just failing silently as the "new portal" does.