Resolved
The issue has been resolved, and the team is continuing to investigate the source of the spikes.
Investigating
From 8:49 PM to 8:54 PM UTC, and 9:11 PM to 9:19 PM UTC, we experienced increased latency of the Eleven Flash v2.5 and Eleven Turbo v2 Text to Speech models.
This also affected a percentage of calls in Conversational AI, leading to crashes.