On September 19th around 9am PDT, we received reports from clients that content lists were sometimes failing to load and otherwise loading slowly on mobile apps, tv apps, and web embeds. During the disruption, which lasted from 8:53am-9:25am PDT, content list services responded to 83% of requests. The other 17% failed.
We determined that a memory issue was at the root of the problem. After rebooting the affected nodes, we immediately implemented a short term fix, then later a long term fix.