We just added the osd_memory_target config option that makes the daemon adjust the bluestore_cache_size dynamically to meet that target. Can we update teh mds and mon to follow the same pattern? For the mds, IIRC we recently added the mempool tracking needed to track cache size usage. We could reuse the OSD strategy of comparing to RSS and adjust the mds cache size up or down in order to stay within a configured envelope...? For the mon, we have at least two caches we can adjust: the rocksdb cache, and mon_osd_cache_size. Should we do the same thing there? My only concern with a mon_target_memory option is that we have to set a default, and the reality is that we want the mon to use more memory for large clusters. Perhaps we could issue a health alert if it looks like the mon is targetting too little memory for the cluster size, and then set the default for something that works well for medium-ish average clusters? For the mgr.. I'm not sure we have anything we can really tune there to control it's memory usage... it will consume what it needs given the clsuter size, which modules are enabled, and so forth. sage
