AI SBCloud Service Degredation
Timeline · 1 update
- identified Jul 31, 2026, 08:00 PM UTC
We have identified a issue impacting access to AI services in South Bend. Engineers are working to restore functionality.
Aunalytics had 8 outages in the last 2 years totaling 48h 45m of downtime — averaging 0.3 incidents per month.
There were 8 Aunalytics outages since April 29, 2026 totaling 48h 45m of downtime. Each is summarised below — incident details, duration, and resolution information.
We have identified a issue impacting access to AI services in South Bend. Engineers are working to restore functionality.
We have identified an issue impacting our internal LLM hosting services. Our team is actively investigating and working to restore normal service as quickly as possible. Additional updates will be provided as more information becomes available.
We have identified an issue affecting one of our storage arrays in our South Bend, Indiana data center. Our team is actively investigating and working to determine the cause. Some customer workloads may experience an impact while the issue is being addressed.
A fiber issue has been identified between our Bellefontaine POP and Plain City POP. There are currently no service interruptions, and all traffic has successfully failed over to the redundant path as designed. We are currently working with our fiber provider to locate the issue and will continue to monitor the situation and provide updates as additional information becomes available.
We identified some missing enrichments on tickets after a backend model issue was identified and fixed.
Cloud Portals in COL are in a disruptive state. We are currently investigating to remediate.
Aunalytics has indentified an issue with the Comcast point to point circuit between KZ and South Bend. The circuit is experiencing packet loss which is dropping routing protocol between the sites. Customers may experience packet loss on ISP or Data Center to Data Center connectivity. We are engaging the carrier now.
Internal GPU Services experienced an outage, and we believe we have identified the underlying issue. The fix is currently being applied.