The current alert trigger for index suggestion is described here: https://docs.atlas.mongodb.com/reference/alert-conditions/#host-alerts

"If the query targeting ratio for a host consistently exceeds 10,000 for a period of 10 minutes, Performance Advisor checks the host for inefficient queries and possible indexes to improve performance."

Firstly we have issues with alerts constantly going on and off every few minutes if the query ratio is fluctuating around the 10k mark.

Secondly, we often do not get alerts for worthwhile index suggestions. We run a multi-tenant platform. We have big clusters with hundreds of databases/tenants and thousands of collections. A huge amount of our queries are optimized and will have a query ratio of 1. Because the query ratio is calculated for the whole cluster, the alert is not triggering often enough.

We need an alert trigger that is not based on the query ratio.