How IT Infrastructure Management Can Prevent Azure Cloud Storms
One of the big attributes of cloud technology is the ability to dynamically adjust resources so that when more capacity is needed, you simply click a few buttons. There’s no need for a lot of meetings or analysis to positively impact system or application performance. You look over a few logs and graphs, determine where the bottleneck is occurring, and add capacity to your IT infrastructure.
Sounds great, right!? Not always...
Unexpected Azure Downtime
Recently, a client of ours who is a software developer opened a support ticket because they could not get their development Azure-hosted virtual server started one Monday morning so that their team could work on coding changes. It took days to get anything back from Microsoft, which is not all that surprising. What was surprising to me was the response we did eventually get from them.
They told us that the US South data center was at capacity and resources were not available to allocate to our client’s server. Their solution was to move it to another data center where more resource availability exists. That would be an easy solution, except that their documentation states that additional resources must be allocated to the server in the source environment to move it.
If there aren’t enough resources to start the VM, how will there be enough resources to move it?
Improving IT Infrastructure: Short-Term Fixes and Long-Term Solutions
After several days of trying, we were able to get the VM started again. However, this meant that our client had several days where their development team was staring at their computer screens accomplishing nothing. So we have scheduled to move their virtual servers to another data center which should address the problem in the short term.
Having inherited those virtual servers and their configurations from a prior support company, Solve iT is taking long-term measures to improve reliability as well as reduce operating costs associated with those Azure objects and overall IT infrastructure.
Cost Management Strategies
Since the use of Azure objects is metered, there are some options to consider when setting up or managing Azure virtual servers. First, the use of a Reserved Instance is a well-documented mechanism. A reserved instance is simply an agreement that says we want to run our server(s) for X number of hours per week and for that commitment, we get a discounted rate on the objects. What that doesn’t address is the metered use of the objects associated with the virtual server(s).
Additionally, there are not very many “experts” to advise on this because the technology is still fairly new. Therefore, we have found through trial and error that scheduling virtual servers to shut down at night when nobody is using them helps reduce these metered expenses when the systems are actually not needed. Mind you, this type of scheduling gets tricky and needs to take into account regular maintenance routines such as applying security updates to the operating system and applications, backups, etc.
These things, along with supplying your own licenses, or what Microsoft Azure labels as Hybrid Licensing, all help to bring costs down significantly for Azure hosting.
The Shift to Cloud Technologies
When faced with a $25-30K capital expenditure upgrade cost per server, and an equipment ordering lead time of 2-3 months, many of our clients have opted to migrate away from on-premise servers by adopting cloud technologies. Most application vendors are now offering cloud-based versions of their software which is greatly reducing the need for on-premise servers. By rolling into the mix many of the security features available in the Microsoft 365 Premium package, IT providers are able to leverage the cloud-based capabilities that replace legacy on-premise Active Directory services at a much more agreeable regular monthly expense.
Addressing Capacity Issues
Circling back to the capacity issue with the Microsoft data center, there seems to be no acceptable answer as to how or why Microsoft chooses systems to shut down when capacity becomes a problem at one of their data centers. But, a Reserved Instance can help ensure that does not happen to your business.
In our research, we uncovered information that indicates Microsoft is having the same supply chain issues the rest of us face when sourcing server hardware. This, I’m told, is the reason for the resource limitations in that data center. It is certainly understandable, but some heads up on that sure would have been welcome for us and for our new client. Ensuring a reliable IT infrastructure is crucial, and understanding these limitations can help in planning better.
FREE TEST DRIVE
Take our tools for a free spin, no commitment required...
(seriously!)
