Thousands of users reported service issues with the Microsoft products on Thursday, a day after a brief outage happened with the same products.

REDMOND, Wash. — Thousands of people reported issues with various Microsoft products on Thursday, just a day after the services experienced a brief outage.

More than 15,000 users reported issues using Microsoft 365 products at around 3 p.m. Eastern, according to the outage-tracking site DownDetector. Another 12,000 people reported issues with Microsoft Outlook, and 500 reported issues with Microsoft Teams.

Microsoft Support teams were investigating the outage, the company said in a social media post. Investigators blamed the issue on the service infrastructure.

“We’ve identified a portion of service infrastructure in North America that is not processing traffic as expected,” the company said. “We’re working to restore the infrastructure to a healthy state to achieve recovery. More information can be found at https://status.cloud.microsoft or under MO1221364 if accessible.”

Nearly two hours after the outage began, investigators said they’ve restored infrastructure to a “healthy state,” but said further load balancing is required to mitigate the outage’s impact.

“We’re directing traffic to alternate infrastructure to achieve recovery,” Microsoft said in a social media post at 4:14 p.m. Eastern. “More information can be found at https://status.cloud.microsoft or under MO1221364 if accessible.”

In the most recent update, Microsoft said it was “rebalancing traffic across all affected infrastructure to ensure the environment enters into a balanced state.” 

“We’re proceeding as quickly as possible and this approach will help us identify any additional actions needed for recovery,” Microsoft said in a social media post. “More info can be found at https://status.cloud.microsoft or under MO1221364 if accessible.”

As of 6:20 p.m. ET, reports on DownDetector were down to 3,000 for Microsoft 365. 

Microsoft resolved a brief outage on Jan. 21, which investigators said was caused by a third-party networking issue. That outage only lasted just over an hour.

This article will be updated with the latest information as it is released.