Open
Description
We have observed certain Rails operations - particularly ActiveSupport related ones - produce an immense amount of notifications within a fairly short amount of time. We should research the concept of enforcing either a per notification event or per Rails library threshold beyond which the agent simply ignores new notifications until the seen tally is reset (likely at harvest time). After a certain amount of notifications has been seen, the value of the additional observed events may decrease or the agent may itself be wasting resources on processing the notifications.