At 11:47 CEST, we lost access to a cloud instance handling our token service. A few minutes after that, the team was looking into it, and the service was restarted at 11:53
We added an extra cloud instance to increase resilience for this service in the future.