Summary:
Despite previous improvements in cache management, the cache server ran out of memory, causing slowdowns and intermittent errors.
Immediate Resolution:
The service recovered automatically, and the disruption was brief.
Next Steps:
After closer inspection of the root cause, we will make further improvements in cache memory management and add more memory to the cache servers next week during monthly scheduled maintenance.