Not a lot happened last week in the world of Azure updates of interest to developers.
One item of note is that you can now apply a service tier setting to a specific namespace in Notification Hub, Azure’s push notification offering.
Until now, you could choose only one service level (free / basic / standard) for all the namespaces in your Azure subscription. That meant that if you have many different messaging needs, you needed to choose the service tier for the worst-case situation. Which is counterproductive in the cloud environment.
For example, maybe your solution had need to send push notifications to a small subset of users rarely, but an ongoing need to send push notification to large segments of your users frequently. You couldn’t group both those namespaces together without putting them both in the standard tier; you would need to create a whole new subscription to use the free tier, and that would make tying related services to that alternate subscription difficult to impossible.
This change will allow you to create namespaces that serve specific functionality and to set an appropriate service level on that namespace.
Preview: Data import / export and scaling in Redis
Maybe I’m doing it wrong but I don’t see a huge benefit to import / export of memory cache. It kind of defeats the purpose, in my view.
I also see limited need to autoscale Redis to the premium service tier; if that’s necessary, maybe your architecture is too reliant on cache and is in serious need of simplicity. But then again, I’m not running websites or webservices with tens of millions of visitors every minute.
Generally speaking, things are in preview for a month or two before they start becoming standard features.
Protect resources from meddling
One thing I do see a great value in is the new resource lock feature, which prevents accidental deletion or reconfiguration of Azure services that you select.
The nice thing about this approach is that is can be broadly applicable across all users within a subscription. For example, I can set everyone to being unable to delete resources, and institute an external process for review of all service deletions, which tend to be harder to recover than misconfigurations.