Problem Statement

If a critical infrastructure component goes down (say a DB cluster becomes unreachable due to network issues), an application will start generating a lot of errors.

When this happens, small organizations which do not have staff available 24/7 can incur in the very unpleasant situation in which this repeated error happens during a holiday or weekend while the staff is off and consumes not only the entire error quota, but also the on-demand spend.

Solution Brainstorm

To avoid this, it would be really great if there was a way we could tell sentry to discard the same occurrence of the same error if there's more occurrences of X times within X hours. For example, I wouldn't want to receive more than 200 instances of an error within 24 hours. By the time I wake up I will find out and intervene.