Incident Report 12/05/2025

Start: Unicorn/2.18.0 (rollout: Feb 27- Mar 17, 2025)

End: May 12, 2025 (migration fixed and applied)

Reach: Applies to flows that were incorrectly migrated and use a scheduler with a daily recurrence set to hour ‘24’.

Impact: high - blocking

Description

Some scheduled flows failed to execute, and associated notifications were not triggered. This issue affected flows using a scheduler configured with a daily recurrence and the hour “24” selected, which were not correctly migrated by our underlying pipeline.

We know how critical scheduled automations and timely alerts are to your workflows, and we sincerely apologize for the disruption.

Our team acted to identify the root cause and implement a fix.

Causes

• When a scheduler is added and set up, its configuration is translated into a CRON expression readable by AWS infrastructure so that it becomes activated.

This issue appears with a specific scheduler configuration. A flow containing a scheduler step with a schedule set to “At regular intervals” with an “At” (hour) containing the “24” option ticked could not be deployed. This configuration generates an invalid CRON expression (0-23, not 24), which is refused by AWS.

• This issue prevents the flow from being deployed correctly; as a result, no executions were triggered, and therefore no notifications were sent.

Resolution

• The function that translates user configurations into CRON jobs was fixed and tested; however, the migration did not run correctly, and this issue went undetected.

Follow-up

• The Mindflow engineering team is committed to ensuring platform reliability and preventing issues like this in the future. As part of our ongoing improvements, we will enhance the notification system to detect and surface internal execution failures—even in cases where misconfigurations prevent flows from deploying or running. Additionally, we are strengthening our quality assurance processes around scheduling logic to catch edge cases earlier in development.