Due to the slow down from the earlier incident in cluster us-central1-d8ff, we had large backlog of tasks which caused parts of the message broker to exhaust resources. We were able to quickly provide more resources to the message broker which allowed the integrations to speed back up.
No data loss occurred during this degradation and the message broker has more permanent resource headroom to avoid this in the future. We apologize for any inconvenience caused by this incident.