Previous incidents

Following a platform update at 07:15 CET, we observed performance slowdowns starting at 08:46 CET.
These slowdowns impacted the handling of inbound calls:

Some calls were not processed.
Callers received a message asking them to call back later.

A full rollback to the previous version was executed at 09:06 CET. Since then, the system has been stable and is under active monitoring.

Our R&D team identified that between 15:14 and 15:46 CET, post-call data extraction was unavailable.
Calls were successfully handled, but extracted information was missing.
The issue was caused by a platform update and has been resolved. The system is now fully operational.
We continue to monitor closely to ensure stability.

Our R&D team identified that at 13:03 CET today, one backend instance became unresponsive. Less than 13% of calls were affected, resulting in unanswered calls. In most cases, callers retried and their calls were successfully handled.
The issue was resolved at 17:54 CET, and the platform is now fully operational. We continue to monitor the situation closely to ensure ongoing stability.