We reduced the load by 3x on a problematic client, hoping this would help mitigate the circumstantial issue in the future.

Issue seem to come from increased usage of our reporting dashboards, we're making scaling adjustements.

We're seeing some dropped requests, we're investigating.