Open
Description
Component(s)
exporter/elasticsearch
Describe the issue you're reporting
Due to nanosecond timestamp resolution, the probability of events having a Stacktrace.count
value > 1 is very low. For the rare cases where it is, we can write multiple identical events. The compression in Elasticsearch will handle this very well.
On the read/query side, things would be come less complex, and random sampling could be applied (requirement is that every event has the same "weight" aka count
). Random sampling can reduce latency for queries.