Service availability was temporarily affected for users whose CI/CD pipelines and platform integrations depend on internally managed credentials. The issue began on May 6, 2026 at 15:01 (UTC-5) and was identified proactively by our engineering team at the time of onset. Service was restored at 15:40 (UTC-5) the same day, resulting in a total window of exposure of approximately 39 minutes (WOE).
During a previously scheduled credentials rotation, a subset of service credentials was not correctly carried forward into the updated secrets configuration. This caused dependent internal processes to lose access to the credentials they required, temporarily affecting the availability of certain platform features and integrations.
Our engineering team initiated a new full credentials rotation cycle, systematically updating credentials across all affected platform components. Service was restored once the update of the affected components was completed. No credential values were exposed as part of this incident.
This incident highlights the importance of ensuring all credentials are fully validated and carried forward before concluding a rotation cycle. We are reinforcing our rotation procedures to include verification steps that confirm all credentials are correctly propagated across every dependent component prior to finalizing the process, reducing the likelihood of a similar issue in the future.
ROTATION_FAILURE < MISSING_TEST < INCOMPLETE_PERSPECTIVE