This incident has been resolved.
Our caching service, which facilitates most read related operations, stopped accepting new items. This slowly increased load on our backend as more load was slowly introduced that is supposed to be dealt with by the cache. We've identified and applied a mitigation for this to improve the cache's availability. We will also be reviewing additional configuration with how our app responds to a loss of the cache in general and applying fixes to further improve it's availability and reliability.
