Based on the Slack report for the support case https://support.mongodb.com/case/00851959, a few other mentions of the similar issue here and there, and recommended collections/indexes sizes in the Atlas doc we probably want Compass to not struggle at least when people are hitting the recommended maximums for those on the m20/30 Atlas clusters.

Currently we are seeing issues when cluster has more than 1k databases or collections in a single database, partially it's just browser struggling to render stuff on the screen, to some extent its views re-rendering even when they don't really need to due to plugins passing all the possible state around, but also it's because we are currently fetching all the data for all the databases and all the data for all the collections in the selected database at once, exhausting the driver connection pool, which means that even if you get through the app lag and manage to open the CRUD screen, you will not be able to fetch the data until all the previous requests in the pool finish.

One way to handle that would be to fetch data only for the databases and collections that you see on the screen and async load it when the items are scrolled into view.