Cosmos is pretty straightforward. There is more complexity, so you just have to be mindful. We had a small issue with making sure that the disaster recovery settings were set up correctly. We found out that there was some type of outage in the middle of the night, but we noticed that the failover didn't run properly. It was because of some configuration that should have been caught earlier, and it wasn't obvious that we got it wrong. There are some infrastructure teams that manage some underlying resources that are related to Cosmos and some of the configurations, but for our specific implementation, we have three developers at most. We usually only need two people for maintaining and managing the solution.