Heathrow, NatWest and Minecraft are among some of the sites and services experiencing problems amid a global Microsoft outage.

Outage tracker Downdetector showed thousands of reports of issues with a number of websites globally on Wednesday.

Microsoft said some users of Microsoft 365, which includes Outlook and Teams, might see delays.

The company’s Azure cloud computing platform, which underpins large parts of the internet, reported a “degradation of some services” at 1600 GMT.

It said this was due to “DNS issues” – the same root cause of the huge Amazon Web Services (AWS) outage last week.

Amazon says AWS is currently operating normally.

Other impacted sites in the UK include supermarket Asda and mobile phone operator O2 – while in the US, people have reported issues accessing the websites of coffee chain Starbucks and retailer Kroger.

Microsoft said business Microsoft 365 customers might see problems.

It said it had found parts of its infrastructure with connectivity issues, and was working to “reroute affected traffic to restore service health”.

It has started a thread on X with updates after some users reported they could not access the service status page.

Meanwhile, business at the Scottish Parliament has been suspended because of technical issues with the parliament’s online voting system.

A senior Scottish Parliament source told BBC News they believe the problems are related to the Microsoft outage.

Microsoft has been contacted for comment.

On its service status page, Azure’s network infrastructure was showing as “critical” in every region in the world.

Exactly how much of the internet is impacted is unclear, but estimates typically put Microsoft Azure at around 20% of the global cloud market.

The firm said it believed the outage was a result of “an inadvertent configuration change”.

In other words, a behind-the-scenes system was changed, with unintended consequences.

Microsoft said it plans on fixing the problem by effectively replacing its service with a recent backup it knows was working properly.

But it could not give an estimate for how long this would take.

The concentration of cloud services into Microsoft, Amazon and Google means an outage like this “can cripple hundreds, if not thousands of applications and systems,” said Dr Saqib Kakvi, from Royal Holloway University.

“Due to cost of hosting web content, economic forces lead to consolidation of resources into a few very large players, but it is effectively putting all our eggs in one of three baskets.”