At the start of the global Covid-19 pandemic in 2020, Microsoft officials took the lead in explaining how the company was trying to avoid cloud computing capacity issues despite rising demand.
Two years later, Microsoft customers are still experiencing capacity issues, due to ongoing supply chain limitations. The difference today is that officials aren’t saying as much publicly about what’s going on. In private, however, they might.
A limited place
A recent report from The Information indicates that capacity issues are affecting Azure data centers in the United States (in Washington State), Europe and Asia due to supply chain issues.
According to the report, more than two dozen Azure data centers around the world are operating with limited server capacity. And this capacity should remain limited at least until the beginning of next year, can we read.
Microsoft customers and partners have thus realized that the cloud is not unlimited. It is indeed limited by the speed at which large enterprises can scale their data centers. Yet two or three years ago, the idea that there could be no more room in the cloud would have seemed crazy. This is now the new normal, not only for Microsoft, but also for AWS and Google.
Cloud capacity quotas
Microsoft has informed some of its customers of the existence of quotas on the capacity of its cloud. It’s no surprise, given that its customer base is largely made up of enterprises, that Microsoft prioritizes its capacity for existing business users, which means newer and pending customers are more likely to come up against quotas and barriers that are difficult to overcome.
“I didn’t think that Azure having capacity issues was news. I thought that thanks to all the jokes on social media, everyone took it for granted that the effects of the pandemic and the freight crisis had crippled the supply of electronics products around the world, ”wrote on his blog Aidan Finn, Microsoft Most Valuable Professional (MVP) and Principal Consultant for Innofactor Norway.
General shortage
Two years ago, Microsoft announced that it was limiting the creation of new data centers due to the shortage of chips and servers. This situation was apparently later rectified. A year ago, the Redmond firm said it was on track to build 50 to 100 new data centers (not data center regions, but actual physical data centers) each year.
But continued supply chain uncertainties have made it difficult to achieve this goal. It’s not just chips and servers that are in short supply. Everything from power supplies to concrete is also hard to find.
Combine these shortages with Microsoft’s decision to prioritize setting up cloud security services when needed to help with the war in Ukraine, and you have a situation where even the best plans Workload Balancing can’t perform miracles when it comes to capacity.
Anticipate needs and expenses
However, Microsoft does not specify to what extent the capacities of Azure are limited. But the software giant acknowledges that existing customers will benefit from the offer first.
“Around the world, we have seen unprecedented cloud growth. Faced with this boom, coupled with macro-trends impacting the entire industry, we have taken steps to accommodate customer capacity increases while accelerating the deployment of servers in our data centers,” said a Microsoft spokesperson. “Our priority remains to ensure business continuity for customers. In addition to managing and planning for growth, we actively perform load balancing where necessary. If it becomes necessary to put capacity restrictions in place, we will first limit testing and internal workloads to prioritize the growth of existing customers,” he adds.
What can be done to limit the impact of these capacity problems?
First, plan your cloud computing needs and expenses at least two to three years in advance. You should discuss with your Microsoft representatives what Azure will be able to handle during this time. According to people I’ve asked about this, you should also keep an eye on quotas and anticipate any scaling needs six months in advance.
While some Microsoft partners advise their customers to consider older VM SKUs as a possible way around capacity issues, others say it’s best to try using the latest SKUs, even if they require newer and perhaps less readily available material.
And if you’re a large enough customer, multi-regional architectures spanning multiple Availability Zones can help with provisioning – provided your business can afford to pay for the service.
You can also optionally disable auto-scaling and reserve VM capacity in advance, per Aidan Finn’s suggestions.
Source: ZDNet.com