Resolved -
This incident has been resolved.
Oct 23, 16:02 PDT
Identified -
We've identified the root cause and are working on the fix to the web app performance issue.
Oct 23, 14:55 PDT
Investigating -
Inference calls are *not* affected.
Model management API (e.g. new deploys, changing auto-scaling settings) are also showing higher than usual latencies.
Oct 23, 14:46 PDT