Microsoft’s Copilot AI service suffered a web outage lasting around 100 minutes, while a separate resilience test disrupted suggested prompts for some Copilot Chat users.
Copilot website hit by Cloudflare Error 1016
The copilot.microsoft.com website became unavailable overnight, with users attempting to access the AI assistant encountering a Cloudflare Error 1016 message.
The disruption lasted from 22:25 UTC on 9 September until 00:05 UTC on 10 September, leaving the Copilot website inaccessible for approximately one hour and 40 minutes.
Microsoft later confirmed that engineers had identified a problem affecting the relevant network flow and implemented a configuration change.
“We’ve diagnosed the issue occurring in the affected network flow and applied a configuration fix, which we’ve confirmed has fully restored copilot.microsoft.com access,” Microsoft said.
The incident affected the web-based Copilot service and was subsequently marked as resolved.
Separate Microsoft 365 Copilot test causes disruption
At around the same time as the website outage, Microsoft’s 365 Copilot team was conducting a resilience test that produced an unintended side effect for some users.
The exercise affected Copilot Chat’s suggested prompts, sometimes known as “suggestion pills”, which normally appear after the chatbot provides a response.
“Users may have been unable to see the suggestion pills after receiving a bot response in Copilot Chat,” Microsoft said.
What are Copilot suggestion pills?
CopilotKit defines a suggestion pill as a “suggested prompt shown to the user (typically as a clickable pill)”.
These prompts are intended to make it easier for users to continue a conversation with an AI assistant by offering possible follow-up questions or actions.
Microsoft stopped the resilience exercise after discovering that it was affecting the feature.
“We’re reviewing our internal drill procedures to better identify the potential for impact to suggestion pills in Copilot Chat prior to beginning the drill, so that we can prevent similar impact in the future,” the company said.
Microsoft reviews resilience testing procedures
Resilience testing is designed to help technology companies determine how their systems behave when components fail or encounter unusual operating conditions. Such exercises can be particularly important for cloud-based platforms relied upon by businesses and organisations.
However, the Copilot incident demonstrated that testing itself can sometimes have unintended consequences for users.
Microsoft’s decision to halt the drill and review its procedures suggests the company will examine how potential customer-facing effects can be identified before similar exercises are conducted.
The resilience-test disruption and the separate website outage were both resolved.
Copilot returns to normal service
The incidents come as Microsoft continues to expand Copilot across its consumer and enterprise software ecosystem, making service reliability increasingly important as more organisations incorporate generative AI tools into everyday workflows.
One observer summed up the disruption with a sceptical reference to wider concerns about artificial intelligence: “And they say AI is going to kill us all when they can’t even keep it running.”
For Copilot users, however, the immediate issue was more straightforward: Microsoft restored access to the website and ended the resilience test responsible for disrupting suggested prompts.
Thomas Hardy is a contributor to OE Mag, covering news, politics, business, technology, sport, entertainment, and lifestyle. He focuses on clear, accurate reporting and useful information that helps readers stay informed about current affairs and developments that matter to them. His work highlights relevant stories, emerging trends, and key issues, presenting them in a balanced, accessible, and reader-friendly way.