We've identified a crash in scheduling mechanism processes which resulted in failed periodic tasks such as exports and case alerts. Once this problem has been discovered, our development team has fixed it right away.
Affected exports stuck in "in progress" state were regenerated. Affected case alerts will be sent out on 02.10.2018 at 23:59 CET.
Survey answering system was not affected. Import processes were not affected. We have no evidence of any kind of data loss.
Affected period of time: 02.10.2018 ~09:38:00 CET - 02.10.2018 ~15:39:00 CET
Total: 6 hours, 1 minute
Why did it happen?
The issue was caused by an unexpected out-of-memory error on the operating host machine.
What are we doing to avoid such issues?
Memory resources were increased for the host machine. Monitoring and alerting procedures will be reviewed to allow for more reactive behavior from the development team in case of production host failures.
We apologize for the inconveniences this incident have caused.