Some users want to start dropping TRACE/DEBUG/INFO logs before WARN/ERROR logs are dropped. As the number of logs/s increases, it might be useful to start increasing the probability that a specific log type is dropped.
Perhaps a set of weights like:
importance_weights:
error: 1.0
warn: 0.9
info: 0.5
debug: 0.1
trace: 0
We can map that to a nice probability curve or something. Not sure what a user would expect / how they'd want to define things.
Some users want to start dropping TRACE/DEBUG/INFO logs before WARN/ERROR logs are dropped. As the number of logs/s increases, it might be useful to start increasing the probability that a specific log type is dropped.
Perhaps a set of weights like:
We can map that to a nice probability curve or something. Not sure what a user would expect / how they'd want to define things.