Rationing the cloud: How coronavirus has created capacity issues

In the face of the current health emergency, Microsoft Azure and Office 365 are inevitably having to scale resources. Andy Powell, Cloud CTO at Jisc, shares advice for dealing with capacity issues

‘These are unprecedented times.’ It’s a sentence you no doubt will have heard a lot in the last few weeks, and one that applies to technology every bit as much as other sectors.

In fact, public cloud services have seen a massive spike in demand since the COVID-19 pandemic hit, and Microsoft have reported that whole countries have gone from not using cloud to deliver teaching at all to 100% cloud-based remote learning – all in a matter of weeks. Collaboration platforms such as Microsoft Teams are probably responsible for the majority of that increased demand.

It’s not surprising, therefore, that we are starting to see the first signs of resourcing issues.

The Register has reported that ‘Azure seems to be full’, and I have seen similar coverage of cloud capacity problems elsewhere.

Move data with caution

On-demand, scalable computing resources, such as virtual machines (VMs), are a popular choice to increase computing capacity through Azure – but with the current surge of cloud custom, users may encounter errors when trying to provision VMs.

For those that do encounter these errors, at the time of writing, Microsoft is suggesting:

• To wait and try again to see if resources get released
• To attempt to recreate the VM in the same region but with different sizing
• To attempt to recreate the VM in another region.

I would suggest, however, that this final option be treated with caution. Although all Azure regions are built to the same level of compliance, there are factors to consider when moving data to a different region, such as the US, and subsequently relying on Privacy Shield frameworks and the EU Model Clause to safeguard this data.

Prioritising emergency services

Cloud providers are prioritising government and emergency services’ use of the cloud. Freeing up capacity for emergency health service use by temporarily moving less critical data to less stressed regions seems to chime well with public interest, but during the current emergency, are there any regions that are ‘less stressed’?

I do not believe that Microsoft will move data out of the region in which customers have placed it. However, what individual customers choose to do is their decision, and the current crisis may tempt some to create resources outside of their normally preferred regions.

In this instance, I would add the following advice to that above:

• Turn off any VMs that are not required for production workloads. This will save capacity for others.
• For production workloads, such as information users can’t afford to lose, make sure VMs aren’t automatically being turned off overnight. There is a danger that these may not be able to be switched back on in the morning.

As with many essentials right now, it’s important that we are all mindful of the resources we are using and the impact that has on others.

Don’t panic!

Capacity issues in Azure will also affect, and be affected by, capacity issues in Office 365. Microsoft have already vastly scaled up this capacity, particularly in response to increased demand for Teams. Customers have very little control over the underlying resource provision of Office 365 since it is delivered as SaaS (software as a service). However, many of the same considerations as outlined above will apply to Office 365.

In all cases, I strongly advise against panic buying: we all know where hoarding toilet roll gets us.