On 3 July between 12:16 UTC and 15:55 UTC, we witnessed an issue in our network between our data centers Equinix AM3 and Equinix AM5.
While trying to find out what the exact issue was, we toggled one of the two interfaces that connect the data centers to each other off and on, which resolved the issue.
We have submitted a case with the vendor to find out what the exact cause of the issue was.