Context:
In Elasticsearch 5 they put a hard limit of 1000 fields per index, while in an ideal world app developers would think about this use case and properly handle it the reality is often that they are under pressure to move fast and deal with issues later.
Basically in addition to creating critical alerts when it does exceed (which we could do based on logs) we should have earlier warnings so that if we get too close those responsible for operations can go back to the app team and say now you need to figure out how to handle this, indices reaching thresholds and errors will start happening otherwise.