Rawpixel - Fotolia


How to live with loss of control with cloud software updates

IT pros don't have the same control over updates and versions in the cloud as they do on premises. Stay ahead of the changes and communicate to staff to lessen any negative impacts.

You come to work Monday morning and login to Microsoft Azure, but nothing looks the same. You check your bookmark, and the page isn't what you expected to find. The portal has undergone a radical change and now you're looking at it like you're a first week intern.

IT pros often handle their tasks quickly because they've done them before. Most IT tasks are muscle memory, with many actions being variations on the same theme. If you don't believe this, put a Windows admin in front of Linux or Mac and watch the struggle.

The recent Azure Portal changes are an example of how cloud software updates can have unforeseen impacts on an IT staff. While Azure is still Azure, certain tasks and functions are in different locations or involve different steps. Adapting to these changes -- depending on their severity -- can take time.

Of course, all software goes through changes -- tools get updated, released in web versions or undergo some alteration over their lifespan. But unlike many other tools IT uses, we don't own the cloud. IT is a customer of the cloud, so changes can occur without out our permission. And if you make a mistake as a result, it can cost you a lot of money.

Unforeseen impacts and lessons learned from cloud software updates

I logged into the Azure Portal to do a few minor tasks after a recent update. Unfortunately, the tasks that were important to me were no longer where I expected them to be. I was able to dig and find them eventually, but I did have to dig. Apparently, what was important to me wasn't as important to the UI designer.

Once I found what I needed I could pin that section and start to build shortcuts again, but customizing my Azure Portal wasn't what I had planned on doing that morning -- and that's the problem. It took time out of my day just to get back to where I was the day before; loss of time to overloaded IT staff never goes over well.

Apparently, what was important to me wasn't as important to the UI designer.

Even worse, the cloud software updates impacted how the portal worked with part of our single sign-on (SSO) system. What should have been a non-issue took several weeks to correct. Admins, applications and developers often rely on these interfaces and services, and changes to them can have far-reaching impacts that come with real and unexpected costs. We often want these new features and functionality, but even in the cloud there are legacy systems -- and people -- that need to keep up.

Updates to cloud APIs, while tested with other clouds and key systems, can run into problems when linked with on-premises systems. This is often caused by the delicate framework that stiches together on-premises systems that aren't always as up to date with different clouds that are more current with their patches and security levels.

Ideally, SSO in the cloud can work pretty easily when you use Azure AD. But if you're still running AD locally, that changes the rules a bit and creates a weak spot in your overall architecture. 

In Microsoft's defense, the change was communicated, and people could even see both portals side-by-side before the change fully occurred. But for your overworked admin, reading change notifications from Microsoft is not high on the priority list and neither is working with the new portal. This isn't Microsoft's fault; it's just a reality in today's IT environment.

What's new in the Azure Portal

The recent Microsoft Azure Portal updates resulted in some changes both in appearance and under the hood.

There were the traditional updates to diagnostics, troubleshooting and health statistics. There was also the addition of Azure Virtual Machine Scale Sets, which allows you to create load-balanced VM groups that can scale up or down depending on load or schedule. There were also improvements to the display naming of your virtual machines -- not a huge technical change, but one that was welcomed by admins.

Advanced Threat Protection for Azure Storage was also officially released as an additional layer of protection against possible storage exploits. You are also seeing improvements in the Azure Kubernetes Service and other developer interfaces, but the real change was in the interface itself. 

The interface moved to a more tabbed browsing environment that is closer to a true web page rather than a thick client. Often vendors will simply do a forklift from a thick client to a web portal that tries to mimic the thick client, but this was different. It's a portal trying to act the way we expect a web page to act, and they nailed it.  Simple things such as a right click to a new window work the way we would expect them to.

Adapt to change

Unfortunately for the cloud admin and application owners, these types of changes are one-sided, and you have to get used to them happening. It's not always pretty or convenient, but when you made the decision to move to the cloud, you gave up that ability to control versions and features.

However, there are a few tips to help survive in this new world. First, read the email notifications from your cloud vendor on changes and patches. You can't control the changes, so your next best option is to be fully aware of them and become the communication point person for your organization.

Take the time to understand the pending shift and how it can impact the applications or services you run. Work directly with those application owners so no one's blindsided. Then, come up with a plan to communicate the changes to your users. Dealing with change isn't always pleasant, but the frustration level will drop considerably if you communicate properly.

This doesn't mean to over-communicate, as people will simply start to ignore those messages, but people will pay attention if you communicate creatively. For example, I've seen major downtime or upgrade notifications delivered with donuts, which was pretty clever.

Also, if you know when the change is going to occur, schedule some time to work with it. Create a meeting for yourself and block off that time to ramp up. It will never happen if you just say you're going to take the time somewhere along the line.

The cloud is all about change and you have to change with it. You can't control how it might change but you can at least be on top of the changes and lead the efforts to ensure your business isn't disrupted when it occurs.

Dig Deeper on Cloud infrastructure design and management

Data Center