At 8:00 a.m. PST, our team began to see an increase in error messages related to inbound silence across inbound and outbound OPI calls placed over IVR and from the web and mobile applications. Around that same time, we started receiving user reports in response to these error messages.
We quickly identified that this was caused by an incident with our PSTN partner, in which an internal network issue significantly degraded call quality or prevented successful connections. Their incident report can be viewed on their status page here: https://status.twilio.com/incidents/5b8wmb5djj4g
While our incident remained open until 10:35 a.m. PDT, we began seeing increases in connection rates and decreases in error messages much earlier, with our PSTN partners taking action to resolve their network issues. By 8:36 a.m. PDT, we had partial recovery and saw full recovery around 9:20 a.m.
Our team chose not to deploy our Delivery Assurance Plan because it relies on the impacted aspects of the PSTN Partner's platform. This approach ensured the highest number of calls were serviced.
Our Development and Engineering Teams are investigating preventive measures to mitigate and improve future responses to such issues. We are also exploring an iteration of our delivery assurance plan that would rely on less overlapping technology with our primary platform.
We are in contact with our PSTN Partner to ensure they are taking the necessary steps to reduce the risk of a similar incident in the future.