we recently had a major outage due to OOM on the mongodb shard nodes . we would suggest help preventing this by providing below features so this can be managed properly going forward 1. Set the memory allocated to the WireTiger Cache 2. Configure more memory for a tier. The current tier is M50 (8 vCPUs, 32GB memory). We would like to be able to allocate more memory for M50 to handle more connections and memory use for workloads. Scaling up to M60 will double the vCPUs to 16 which we do not want. Perhaps offer tier with high memory usage for same number of vCPUs. We discussed this with the support engineer so you can get more details from him. 3.Ability to cap the number of connections on the to prevent a connection storm