> Nothing made any sense until we noticed the controller microservice's memory consumption. A service that should be using perhaps a few hundred megabytes at mo

> Nothing made any sense until we noticed the controller microservice's memory consumption. A service that should be using perhaps a few hundred megabytes at most was using gigabytes and growing... and growing... and growing... and growing... Not identifying this until many hours after symptoms were impacting users sounds like a pretty big monitoring blind spot.