Right amount of indexes to use on a big table?

[Niheel]

One of my applications collects a lot of data and is outputting it to a group of analysts that view it in all sorts of ways. To speed up things we've been using a lot of indexes on our main tables.

At some point, i wonder what kind of problems we can run into with all these indexes we've created.

What's the a good number of indexes for tables with over 10 million rows? Do the number of indexes have an impact on queries or even inserts?

What are some solutions to get by these scaling issues?

[mjhennig]

One cannot simply count the number of columns, rows or fields and calculate a good number of indexes. It depends on, well, many things - but primarily the type(s) of your data.
In my opinion, one can neglect the "maintenance overhead" of indexes on insert, update and delete operations. Some of the databases I'm operating on have tables with over 20 million entries (PostgreSQL, MySQL/InnoDB); but even concurrent modifications aren't considered slow by anyone.