On August 26th, between 05:16 UTC and 16:26 UTC, some ODC users experienced delays in the execution of timers. Affected timers remained pending instead of running as scheduled, and subsequent runs did not start. Users in US East / Virginia and Asia Pacific / Singapore were affected. Applications themselves remained available throughout.
The issue was caused by a recent platform update that changed how a timer's status is recorded once it has run. Under that update, a timer whose status needed to be recorded more than once in sequence failed to update correctly, leaving it pending and blocking later scheduled executions. To restore service, our team returned the affected part of the platform to its previous version, which stopped the issue recurring for affected users. A permanent correction was then developed, validated and deployed across the platform, and we confirmed that timers were executing normally before closing the incident.
We are reviewing our monitoring of timer execution so that issues of this kind are detected automatically rather than through customer reports. We apologise for any disruption caused and thank you for your patience.