At approximately 12:20pm on the 25th of January an auto-heal rule caused our OData instance to recycle at high memory usage, which is standard practice to prevent performance issues. The auto-heal rule was then triggered 15 minutes later due to slow requests, which then caused the orchestrated integration services to stop responding, which impacted order confirmation and integration syncs for workspaces connected to Unleashed.
The Orchestrated Integration services were restored by the monitoring and auto-heal rules, restoring regular service within 15 minutes of the initial crash, at which point integration syncs were functioning correctly and orders were able to confirm through. The OData auto-heal has been temporarily disabled while we perform further testing.