At around 16:41PM we started receiving error alerts indicating that screenshots and other captured data (downloaded files, network logs) could not be saved to our object storage backend. At around 17:11 we identified the cause and rolled out a fix by 17:18. While screenshots were not stored for some steps for jobs running within this period, other execution results were still collected, and executions continued uninterrupted.

We recently changed the authentication mechanism that our execution backend uses to upload screenshots and other information during execution. The team released the latest platform to production today which included this change. However, our release automation did not make the necessary infrastructure changes at the same time which led to the failed authentication events.